CA1341389C - Human granulocyte colony stimulating factor - Google Patents

Abstract

The invention relates to a gene coding for a polypeptide having a human granulocyte colony stimulat-ing factor (human G-CSF) activity, a recombinant vector containing said gene, a transformant containing said vector, and a process for producing a polypeptide or glycoprotein having the G-CSF activity. Human G-CSF
are very useful as an active component of infection protective agents, pharmaceutical compositions for the treatment of leukopenia and pharmaceutical compositions for promoting the recovery of hemopoietic capacity. A
process for producing a large amount of human G-CSF by means of gene recombinating technique is disclosed.

Description

1 3~r1 38g _1_ HUMAN GRANULOCYTE COLONY STIMULATTNG FACTOR
The present invention relates to a human granulocyte colony stimulating factor. More particularly, the present invention relates to a gene coding for a polypeptide having the activity of a colony stimulating factor (hereinafter abbreviated as CSF) which is a specific stimulating factor necessary for the principal purpose of forming colonies of human granulocytic cells. The present invention also relates to a recombinant vector inserted said gene, a transformant containing said vector, a polypeptide or glyco-protein having the CSF activity as produced from said trans-formant, and a process for producing a polypeptide or glyco-protein having the CSF activity.
When bone marrow cells as target cells and kidney cells or fetal cells were cultured by the double-layer soft agar cultivation method, with the bone marrow cells being in the upper layer and the kidney or fetal cells in the lower layer, part of the cells in the upper layer grew and diff er-entiated to form colonies~~of neutrophilic granulocytes (hereunder simply referred to as granulocytes) or monocytic macrophages. This observation has led to the assumption of the presence in vivo of factors which promote the formation of colonies [Pluznik and Sach, J. Cell. Comp. Physiol., 66, 319 (1965): and Bradley and Metcalf, Aust. J. Exp. Biol.
Med. Sci., 44, 287 (1966)].
These factors which are collectively referred to as CSF are known to be produced by cells, such as T-cells, monocytic macrophages, fibroblasts and endothelial cells, which normally are distributed extensively in vivo. Among subclasses of CSF are included: granulocyte-monocytic macrophage CSF (abbreviated as GM-CSF) which act on the stem' cells of granulocytes or monocyte macrophages in such a manner that they stimulate the growth of such stem cells and induce their differentiation to form colonies of granulo-cytes or monocytic macrophages monocytic macrophage CSF
(abbreviated as M-CSF) which is principally capable of 1 3 41 38g -~_ forming colonies of macrocytic macrophages; multipotent CSF
(abbreviated as multi-CSF) which acts on less differentiated multipotent stem cells; and granulocyte CSF (abbreviated as G-CSF) of the type contemplated by the present invention which is principally capable of forming granulocytic colo-nies. It has recently been held that the stages of differ-entiation of target cells differ from one subclass of CSF to another [Asano, Taisha - Metabolism and Disease, 22, 249 (1985); and Yunis et al., "Growth and Maturation Factors", edited by Guroff, John Wiley & Sons. NY, vol. 1, 209 (1983)].
Therefore, purifying the individual CSF subclasses and making a closer study of their chemical and biological properties are very important for the purpose of estimating the hematopoietic mechanisms and analyzing the patho-morphological aspects of various hematological diseases.
The biological actions of G-CSF that are drawing increasing attention of researchers are their capabilities of inducing the differentiation of bone marrow leukemic cells and enhancing the functions of mature granulocytes, and much promise has been held in the potential clinical utility of G-CSF in the fields of tXeating and preventing leukemia.
The attempts heretof ore made to isolate and purify G-CSF are based on the method of cell cultivation wherein G-CSF is isolated from the supernatant of a cell culture, but homogeneous G-CSF has yet to be produced in large quan-tities by this method because G-CSF can only be produced in low concentration and complex purification procedures are required to obtain a trace amount of G-CSF from a large volume of culture solution. .Therefore, it has been strongly desired to achieve mass production of G-CSF by recombinant DNA technology.
One object of the present invention is to provide a gene encoding a polypeptide having the human G-CSF activity.
Another object of the present invention is to provide a recombinant vector incorporating said gene.
Still another object of the present invention is to provide a transformant which has been produced by 1 ~ ~+1 389 transforming a host with said recombinant vector, and a polypeptide or glycoprotein which is produced by said transf ormant .
A further object of the present invention is to provide a process for producing a polypeptide or glycoprotein having the human G-CSF activity.
Fig. 1 shows the sequences of three different probes, IWQ, A and LC;
Fig. 2 shows the nucleotide sequence of a pHCS-1 insert;
Fig. 3(A) shows the nucleotide sequence of a cDNA
insert in pBRG4;
Fig. 3(B) (I) shows the amino acid sequence of a human G-CSF precursor as deduced from pBRG4 cDNA;
Fig. 3(B) (II) shows the amino acid sequence of human mature G-CSF as deduced from pBRG4 cDNA;
Fig. 4(A) shows the nucleotide sequence of a cDNA
insert in pBRV2;
Fig. 4(B) (I) shows the amino acid sequence of a human G-CSF precursor as deduced from pBRV2 cDNA;
Fig. 4(B) (II) shows the amino acid sequence of human mature G-CSF as deduced from pBRV2 cDNA;
Fig. 5 shows the nucleotide sequence of a human chromosomal gene coding for human G-CSF;
Fig. 6 shows the restriction enzyme cleavage sites of pBRG4- or pBRV2-derived human G-CSF cDNA;
Fig. 7 shows the restriction enzyme cleavage sites of the human chromosomal gene coding far human G-CSF;
Fig. 8 is a partial presentation of the process for preparing a tac promoter-containing vector (+VSE line):
Fig. 9 is a presentation of the process for preparing a PL promoter-containing vector (+VSE line);
Fig. 10 is a presentation of the process for prepar-ing a trp promoter-containing vector (+VSE line);
Fig. 11 is a partial presentation of the process for preparing a tac promoter-containing vector (-VSE line);

1 3 ~~ 3a9 Fig. 12 is a presentation of the process for prepar-ing a PL promoter-containing vector (-VSE line);
Fig. 13 is a presentation of the process for prepar-ing a trp promoter-containing vector (--VSE line);
Fig. 14 shows schematically the structure of pHGA410;
Fig. 15 is a presentation of the processes for con-structing expression recombinant vectors, pTN-G4, pTN-G4VAa and pTN-G4VA~;
Figs. 16a and 16b show two processes for constructing pHGG4-dhfr;
Fig. 16c shows the processes for constructing pG4DR1 and pG4DR2;
Fig. 17 shows schematically the structure of pHGV2;
Fig. 18 is a presentation of the processes for con-structing expression recombinant vectors, pTN-V2, pTN-VAa and pTN-VAS;
Figs. 19a and 19b show two processes for constructing an expression recombinant vector pHGV2-dhfr.
Fig. 19c shows the processes for constructing pV2DRl~
and pV2DR2;
Fig. 20 shows schematically the structure of pMLCE3a;
Fig. 21 shows schematically the structure of pTNCE3a;
and Fig. 22 shows schematically the structures of pD26SVCE3a and pDRCE3a.
The gene coding for a polypeptide having the human G-CSF activity according to the present invention is a DNA
(cDNA) which is complementary to the messenger RNA (mRNA) that is obtained as 15 - 17S factions by sucrose density gradient centrifugation and which codes for a polypeptide having the human G-CSF activity.
The present inventors obtained two lines of this cDNA.
The cDNA of one line has all or part of a gene coding for the polypeptide I or II shown in Fig. 3(B)» More speci-fically, this cDNA has the nucleotide sequence delineated by ATG at 32 - 34 nucleotide positions from 5'-terminus [see Fig. 3(A)] and ,.CC at 650 - 652 nucleotide positions, or by ACC at 122 - 124 positions and CCC at 650 - 652 positions.
Alternatively, the cDNA has the nucleotide sequence shown in Fig. 3(A) or a part thereof. The cDNA of this line is hereinafter referred to as cDNA (+VSE).
The cDNA of the other line has all or part of a gene coding for the polypeptide I or II shown in Fig. 4(B). More specifically, this cDNA has the nucleotide sequence deline-ated by ATG at 31 - 33 nucleotide positions from 5'-terminus [see Fig. 4(A)] and CCC at 640 - 642 nucleotide positions, or by ACC at 121 - 123 positions and CCC at 640 - 642 posi-tions. Alternatively, this cDNA may have the nucleotide sequence shown in Fig. 4(A) or a part thereof. The cDNA of this line is hereinafter referred to as cDNA (-VSE).
The gene described above may be obtained by the following procesures: a mRNA coded G-CSF is first prepared from mammalian animal cells or other host cells having the ability to produce a polypeptide having the G-CSF activity;
the mRNA is then converted to a double-stranded cDNA by any of the known methods; a set of recombinants containing this cDNA (the set is hereunder referred to as a cDNA library) is subsequently subjected to screening by known procedures.
The gene of the present invention also includes a human chromosomal gene coding for a polypeptide having the human G-CSF activity. This human chromosomal gene contains l5 a nucleotide sequence that takes part in transcriptional control and it also contains all or part of the nucleotide sequence shown in Fig. 5.
A chromosomal gene may be obtained by first preparing from human cells a set of recombinants containing a human chromosomal gene (the set is hereunder referred to as a human chromosomal gene library), then subjecting said human chromosomal gene library to screening by known procedures.
The human chromosomal gene may be supplied from any type of human cells such as cells extracted from the liver or kidney or cultured cells such as tumor cells. A human chromosomal gene library may be prepared from human cells by any of the known methods [see Maniatis et al., Cell, 15, 687 (1978); and Maniatis et al., Molecular Cloning, Cold Spring ~ 3 41 38g Harbor Laboratory, p. 269 ff. (1982)], which are illustrated below:
extract a human chromosomal DNA from such sources as human fetal liver with phenol or other appropriate chemicals;
digest the extracted DNA partially or completely with an appropriate restriction enzyme to obtain a DNA fragment of an appropriate length: insert the DNA fragment into a a-phage vector DNA fragment with a T4 DNA ligase or other appropriate ligases, with a linker containing the restric-tion site for an appropriate enzyme such as EcoRI being optionally attached; subsequently, obtain a-phage particles by in vitro packaging method and transform host cells such as E. coli with the resulting a-phage particles.
Examples of the a-phage usable as the vector in the above procedures include Charon 4A and EMBL-3 and EMBL-4.
The mammalian cell which may be used as a source of mRNA supply is a human oral cavity cancer-derived cell strain, CHU-2 (deposited at Collection Nationale de Cultures de Microorganismes, or C.N.C.M., under Accession Number I-483). It should however be understood that in place of such tumor cell strains, cells that can be separated from mammals or any other appropriate established cell strains may be employed. Preparation of mRNA may be achieved by one of the methods that have already been proposed for cloning the genes of several other physiologically active proteins:
for example, the whole RNA is first obtained by treatments with a surfactant and phenol in the presence of a ribo-nuclease inhibitor such as a vanadyl-ribonucleoside complex [see Berger and Birkenmeier, Biochemistry, 18, 5143 (1979)]
or by CsCl density gradient centrifugation following treatment with guanidine thiocyanate [see Chirgwin et al., Biochemistry, 18, 5294 (1979)], then poly(A+) RNA (mRNA) is obtained by subjecting the whole RNA to batch adsorp-tion or affinity column chromatography on oligo(dT)-cellulose or poly-U-Sepharose with Sepharose 2B used as a carrier. The poly(A+) RNA may be further fractionated by an appropriate method such as sucrose density gradient centrif-ugation. The ability of thus obtained mRNA to code for a polypeptide having the G-CSF activity may be confirmed by several methods; for example, the mRNA is translated into a protein and its physiological activities are checked; alter-natively, the identity of that protein is determined with the aid of an anti-G-CSF antibody. More specifically, mRNA
is injected into oocytes of Xenopus laevis for effecting translation [see Gurdon et al., Nature, 233, 177 (1972)], or translational reactions may be performed with rabbit reticu-locytes or wheat germs [Schleif and wensink, "Practical Methods in Molecular Biology", Springer-Verlag, NY (1981)].
The G-CSF activity may be assayed by applying the soft agar cultivation method using bone marrow cells, and techniques for performing this method have been reviewed [Metcalf, "Hemopoietic Colonies", Springer-Verlag, Berlin, Heidelberg, NY (1977)].
A single-stranded cDNA is synthesized with the so obtained mRNA being used as a template; a double-stranded cDNA is synthesized from this single-stranded cDNA; and the double-stranded cDNA is inserted into an appropriate vector DNA to form a recombinant.~plasmid. This recombinant plasmid may be used to transforzn~a suitable host, say Escherichia coli, so as to obtain a group of DNAs in the transformants (cDNA library).
A double-stranded cDNA may be obtained from the mRNA
by one of the following two methods: the mRNA is treated with a reverse transcriptase with oligo(dT) which is comple-mentary to the poly(A)-chain at 3'-terminus being used as a primer; or an oligonucleotide that corresponds to part of the amino acid sequence of G -CSF protein is synthesized, and a cDNA which is complementary to the mRNA is synthesized by treatment with a reverse transcriptase with the synthesized oligonucleotide being used as a primer. A double-stranded cDNA may also be obtained by the following methods: mRNA
is decomposed and removed by treatment with an alkali and the resulting single-stranded cDNA is treated first with a reverse transcriptase or DNA polymerase I (e. g. Klenow fragment), then with S1 nuclease; alternatively, the mRNA
may be directly treated with RNase H and DNA polymerase _8_ (e. g. E.-coli polymerase I). For more information, see, Maniatis et al., "Molecular Cloning", Cold Spring Harbor Laboratory (1982); and Gubler and Hoffman, Gene, 2~, 263 (1983).
The so obtained double-stranded cDNA is inserted into an appropriate vector such as, for example, one of the EK-type plasmid vectors typified by pSC101, pDF4l, ColEl, pM89, pBR322, pBR327 and pACYCl, or one of the phage vectors typified by agt, ac, agtl0 and agtWES, and thereafter, the recombinant vector is used to transform a strain of E. coli (e. g. X1776, HB101, DH1 or C600) so as to obtain a cDNA
library (see, for example, "Molecular cloning", ibid.) The double-stranded cDNA may be joined to a vector by the following procedures: a terminus of the cDNA is Provided with a joinable end by attachment of an appropriate chemically synthesized DNA fragmentp and a vector DNA which has been cleaved with a restriction enzyme is joined to said cDNA by treatment with a T4 phage DNA ligase in the presence of ATP. Alternatively, dC, dG-chains (or dT, dA-chains) are attached, respectively, to the double-stranded cDNA and a vector DNA which has been cleaved with a restriction enzyme, and a solution containing both DNAs is annealed (see "Molecular Cloning", ibid.) A host cell may be transformed by the so obtained recombinant DNA by any of the known methods. If the host cell is E. coli, the method detailed by Hanahan [J. Mol.
Biol., 166, 557 (1983)] may be employed, wherein the recom-binant DNA is added to a competent cell prepared in the presence of CaCl2, MgCl2 or RbCl.
Screening for the cells harboring the desired gene may be performed by several methods which include: the plus-minus method employed in the cloning of interferon cDNA [Taniguchi et al., Proc. Jpn. Acad., 55, Ser. B., 464 (1979)], the hybridization-translation assay method [Nagata et al., Nature, 284, 316 (1980)], and the colony or plaque hybridization method using an oligonucleotide probe which is chemically synthesized on the basis of the amino acid sequence of the protein having the human G-CSF activity _g_ [Wallace et al., Nucleic Acids Res., 9, 879 (1981); and Benton & Davis, Science. 196, 180 (1977)].
The fragment harboring the thus cloned gene coding for the polypeptide having the human G-CSF activity may be re-inserted in an appropriate vector DNA for the purpose of transforming other prokaryotic or eukaryotic host cells.
By introducing an appropriate promoter and an expression-associated sequence into the vector, the gene can be expressed in an individual host cell.
Illustrative prokaryotic host cells include Escherichia coli, Bacillus subtilis, and Bacillus thermophilus. The gene of interest may be expressed within these host cells by transforming them with a replicon (i.e.
a plasmid vector harboring an origin and regulator sequence) which is derived from a host-compatible species. A desir-able vector is one having a sequence capable of providing the transformed cell with selectivity for expressed trait (phenotype).
To take an example, E. coli may be transformed with pBR322 which is a vector capable of replication in E. coli [see Bolivar, Gene, 2, 95 (1975)]. This vector contains both ampicillin- and tetracycline-resistance genes and either one of the properties may be used to identify the transformed cell. Examples of the promoter that is neces-sary for genetic expression in prokaryotic hosts include the promoter of the ~-lactamase gene [Chang et al., Nature, 275, 615 (1978)], the lactose promoter [see Goeddel et al., Nature, 281, 544 (1979)] and the tryptophan promoter [see Goeddel et al., Nucleic Acid Res., ~, 4057 (1980)] and so on.
Any of these promoters may be employed in the production of a polypeptide having the human G-CSf activity according to the present invention.
A eukaryotic microorganism such as Saccharomyces cerevisiae may be used as a host cell and transformed by a vector such as plasmid yRp7 [see Stinchcomb et al., Nature, 282, 39 (1979)]. This plasmid has the TRP1 gene as a selection marker for yeast strains lacking the ability to produce tryptophan, so the transformants can be selected by f -lo-performing growth in the absence of tryptophan. Examples of the promoter that can be utilized for gene expression include an acidic phosphatase gene promoter [Miyanohara et al.. Proc. Natl. Acad. Sci., USA, 80, 1 (1983)] and an alcohol dehydrogenase gene promoter [Valenzuela et al., Nature, 2~8, 347 (1982)].
The host cell may also be derived from mammalian cells such as COS cells, Chinese hamster ovary (CHO) cells, .:-127 cells and Hela cells. An illustrative vector that may be used to transform these cells is pSV2-gpt [see Mulligan and Berg; Proc. Natl. Acad. Sci., USA, 78, 2072 (1981)].
The vectors used to transform these cells contain origin, selection marker, a promoter preceding in position the gene to be expressed, RNA splicing site, palyadenylation signal, etc.
Illustrative promoters that may be used for gene expression in mammalian cells include the promoters of a retrovirus, polyoma virus, adenovirus, simian virus 40 (SV40), etc. If the promoter of SV40 is used, the desired gene expression may be readily achieved in accordance with the method of Mulligan et al, described in Nature, 277, 108 (1979) .
Illustrative origins that can be used include those derived from SV40, polyoma virus, adenovirus, bovine papil-15 loma virus (BPV), etc. Illustrative selection markers that can be used include the phosphotransferase APH (3') II or I
(neo) gene, thymidine kinase (TK) gene, E. coli xanthine-guanine phosphoribosyltransferase (Ecogpt) gene, dihydro-folate reductase (DHFR) gene, etc.
In order to obtain polypeptides having the human G-CSF activity from the above listed host-vector systems, the following procedures may be used: the gene coding for the peptide having the human G-CSF activity is inserted at a suitable site in one of the vectors mentioned above; the host cell is transformed with the resulting recombinant DNA;
and the obtained transformants are cultured. The desired polypeptide may be isolated and purified from the cell or culture solution by any one of the known techniques, i 3 41 389 Eukaryotic genes are generally held to exhibit poly-morphysm as is known for the case of the human interferon gene [see Nishi et al., J. Biochem., 97, 153 (1985)] and this phenomenon may cause substitution of one or more amino acids or a change in the nucleotide seguence but no change in the amino acid sequence at all.
The G-CSF activity may also be possessed by a poly-peptide which is deficient of one or mare of the amino acids in the amino acid sequence shown in Fig. 3(B) or 4(B) or which has such amino acids added thereto, or a polypeptide which has one or more of these amino acids replaced by one or more amino acids. It is also known that a polypeptide obtained by converting each of the cysteine colons in the human interleukin-2 (IL-2) gene to a serine colon has the activity of interleukin-2 [Wang et al., Science, 224, 1431 (1984)]. Therefore, so long as the polypeptides, either naturally occurring or chemically synthesized, have the human G-CSF activity, all of the genes that code for these polypeptides, recombinant vectors containing these genes, transformants obtained by such recombinant vectors, and the polypeptides or glycoproteins that are obtained by cultivat-ing such transformants are included within the scope of the present invention.
Hereunder outlined are the processes for producing the gene of the present invention coding for a polypeptide having the human G-CSF activity, a recombinant vector having said gene and a transformant having this recombinant vector, and a polypeptide or glycoprotein having the human G-CSF
activity expressed in this transformant.
(1) Probe preparation A homogeneous human CSF protein was purified from the supernatant of a culture of a tumor cell line, CHU-2, and its amino acid sequence from the N terminus was determined.
Fragments were obtained by decomposition with bromocyan and treatment with trypsin and the amino sequences of these fragments were also determined [Example 3(i), (ii) and (iii) ] .

-12_ From the determined amino acid sequences, three nucleotide probes, (A), (LC) and (IWQ), having the sequences shown in Fig. 1 were synthesized (Example 4). Probe (A) was of the mixed type composed of 14 successive nucleotides.
Probe (IWQ) was composed of 30 successive nucleotides with deoxyinosine and was a probe of the type used in the cloning of the human cholecystokinin gene [Takahashi et al., Proc.
Natl. Acid. Sci., USA, 82, 1931 (1985)]. Probe (LC) was a 24-nucleotide probe that was synthesized from the nucle-otides at 32 - 39 positions from the N terminus of the amino acid sequence shown in Example 3(i) on the basis of the nucleotide sequence shown in Fig. 3.
Chemical synthesis of nucleotides can be achieved by applying the improved phosphotriester method to the solid Phase method and has been reviewed by Narang [Tetrahedron, 39. 3-22 (1983)].
.Probes based on amino acid sequences at positions other than those in the above-mentioned probes may also be used.

(2) Construction of cDNA library CHU-2 cells were ~iomogenized after addition of a guanidine thiocyanate solution and the total RNA was obtained by CsCl density gradient centrifugation.
Poly(A+) RNA was isolated from the total RNA by column chromatography on oligo(dT)-cellulose. Thereafter, a single-stranded cDNA was synthesized with a reverse tran-scriptase, and RNase H and ~, cali DNA polymerise I were added to obtain a double-stranded cDNA. A dC chain was attached to the obtained double-stranded cDNA, which was joined to a vector, pBR322, to which a dG chain had been attached at the Pst I cleavage site. The resulting recom-binant DNA was used to transform a strain of E. coli, X1776, and a pBR322-line cDNA library was constructed (Examples 5 and 6).
In a similar manner, the double-stranded cDNA was joined to the ~gtl0 vector with the EcoRI linker and ~-phage line cDNA library was constructed (Example 7).

1 341 3~9 (3) Screening Recombinants derived from the pBR322-line cDNA
library were fixed on Whatmann 541 filter paper and a single clone could be selected by colony hybridization with 32P-labelled probe (IWQ). Further study with the Southern blotting method [Southern, J. Mol. Biol., 98, 503 (1975)]
showed that this clone also hybridized with probe (A). The nucleotide sequence of this clone was determined by the dideoxy method [Sanger, Science, 21~, 1205 (1981)].
The nucleotide sequence of the obtained cDNA insert is shown in Fig. 2, from which one can see that this insert consisted of 308 base pairs including probes (TWQ) and (A), and had an open reading frame cading for 83 amino acids containing the amino acid sequence sh own in Example 3(iii).
The pBR322 derived plasmid containing these 308 base pairs is hereunder referred to as pHCS-1 (Example 8).
A DNA fragment containing the 308 base pairs obtained from pHCS-1 was radialabelled by the nick translation method (see Molecular Cloning, ibid.) and, with this fragment used as a probe, the agtl0-derived cDNA library was screened by plaque hybridization [Benton and Davis, Science, 196, 180 (1977)] to obtain five clones. The nucleotide sequence of a clone which was believed to contain cDNA was determined by the same method as described above [Fig. 3(A)].
As shown in Fig. 3(A), this cDNA insert had a single large open reading frame.
The amino acid sequence encoded by this cDNA can be deduced as shown in Fig. 3(A).
Comparison with the N-terminal amino acid sequence of G-CSF protein shown in Example 3(i) revealed that this cDNA
contained a nucleotide sequence which corresponded to both a signal peptide encoded by 90 base pairs starting with the ATG sequence at 32 - 34 nucleotide positions from 5'-terminus and ending with the GCC sequence at 119 - 121 positions, and a mature G-CSF polypeptide encoded by 531 base pairs starting with the ACC sequence at 122 - 124 positions and ending with the CCC sequence at 650 - 652 positions. Therefore, the polypeptide of the amino acid sequence I shown in Fig. 3(B) was composed of 207 amino acids and its molecular weight was calculated as 22292.6?
daltons. The polypeptide of the amino acid sequence II was composed of 177 amino acids and its molecular weight was calculated as 18986.74 daltons (Example 9).
It should be noted that the ATG at 32 - 34 positions or at 68 - 70 positions can also be considered to be the protein initiation site. Escherichia coli strain X1776 harboring pBR322 which had this cDNA (+VSE) at the EcoRI
cleavage site has been deposited with the Fermentation Research Institute, the Agency of Industrial Science and Technology (FERM BP-954).
Fig. 6 shows the restriction enzyme cleavage sites of the gene .
This cDNA was joined to pBR327 CSoberon et al., Gene, 9, 287 (1980)] at the EcoRI site and the resulting plasmid is hereunder referred to as pBRG4.
The thus obtained pBRG4 was treated with a restric-tion enzyme, EcoRI, to obtain a DNA fragment containing cDNA
of about 1500 base pairs. This fragment was radiolabelled by the nick translation method (see Molecular Cloning, ibid.) and, with this radiolabelled DNA fragment being used as a probe, the agtl0-derived cDNA library was screened once again by plaque hybridization (see Benton and Davis, ibid.) In this plaque hybridization, two sheets of a-phage DNA
fixed nitrocellulose filter paper were prepared; one of these sheets was used for the above-mentioned plaque hybrid-ization and another one was subjected to plaque hybridiza-tion with the already described probe (LC). The phages which turned positive for both probes were selected. A
clone which has a "full-length" cDNA was selected and the nucleotide sequence of the cDNA insert as determined by the dideoxy method is shown in Fig. 4(A).
This cDNA had a single large open reading frame and the amino acid sequence that would be encoded by this cDNA
was deduced as shown in Fig. 4(A).
Comparison with the N-terminal amino acid sequence of G-CSF protein shown in Example 3(i) revealed that this a ~ 4~ ~e9 cDNA contained a nucleotide sequence which corresponded to both a signal peptide encoded by 90 base pairs starting with the ATG sequence at 31 - 33 nucleotide positions from 5'-terminus and ending with the GCC sequence at 118 - 120 positions, and a mature G-CSF polypeptide encoded by 522 base pairs starting with the ACC sequence at 121 - 123 positions and ending with the CCC sequence at 640 - 642 positions. Therefore, the polypeptide of the amino acid sequence I shown in Fig. 4(B) was composed of 204 amino acids and its molecular weight was calculated as 21977.35 daltons. The polypeptide of the amino acid sequence II was composed of 174 amino acids and its molecular weight was calculated as 18671.42 daltons (Example 10).
It should be noted that the ATG at 58 - 60 positions or at 67 - 69 positions can also be considered to be the protein initiation site.
Escherichia coli strain X1776 harboring pBR322 which had this cDNA (-VSE) at the EcoRI cleavage site has been deposited with the Fermentation Research Institute, the Agency of Industrial Science and Technology (FERM BP-955).
Fig. 6 shows the restriction enzyme cleavage sites of the gene. This cDNA was joined to pBR327 at the EcoRI site to form a plasmid which is hereunder referred to as pBRV2.

(4) Screening a human chromosomal gene library A human chromosomal gene library that was prepared in accordance with the procedures described by Maniatis et al.
(Molecular Cloning, ibid.) was subjected to screening with the pHCS-1 shown above. Probes that may be employed in screening include: a pHCS-1-derived 308-by DNA fragment, a pBRG4-derived ca. 1500-by DNA fragment, a pBRV2-derived ca. 1500-by DNA fragment, a DNA fragment of an appropriate length containing part of one or more of these DNA frag-ments, as well as the aforementioned oligonucleotide probes [i.e., (IWQ), (A) and (IC)]. The case of using the pHCS-1 DNA fragment is hereunder described.
This DNA fragment was radiolabelled with 32P in accordance with the nicl~ translation method [see Roop et al., Cell, 15, 431 (1978)]. With the resulting 32P-labelled fragment used as a probe, the human chromosomal gene library was subjected to screening by plaque hydridization (see Benton and Davis, ibid.) so as to obtain ten-odd clones.
After recovering DNA from the clones, a restriction enzyme map was prepared by known procedures [Fritsch et al., Cell, 19, 959 (1980)].
With the same DNA probe being used, Southern blotting (see Southern, ibid.) was conducted and it was found that a DNA fragment of about 4 kb that was cut out with EcoRI and XhoI could potentially contain a region for encoding the human G-CSF polypeptide. Therefore, the ca. 4-kb DNA frag-ment was inserted into pBR327 at the EcoRI site using an EcoRI linker so as to obtain pBRCE3a. With this plasmid being used as a base sequencing DNA, the nucleotide sequence of the ca. 3-kb portion of that ca. 4-kb DNA fragment was determined by the dideoxy method. As a result, said DNA
fragment was found to be a gene coding for the human G-CSF
polypeptide (Fig. 5).
E. coli strain X1776 harboring pBRCE3s (i.e. the plasmid pBR327 having said ca. 4-kb DNA fragment inserted into the EcoRI site) has been deposited with the Fermenta-tion Research Institute, the Agency of Industrial Science and Technology (FERM BP-956).
Comparison between the pBRG4 cDNA insert shown in Fig. 3 and the pBRV2 cDNA insert shown in Fig. 4 revealed that the DNA fragment under discussion contained five exon portions and that it coded for the amino acid sequences deduced from pBRG4 and pBRV2.
Fig. 7 shows the restriction enzyme cleavage sites of the obtained gene.
This DNA fragment contained the chromosomal gene of human G-CSF, or the preceding region to be transcribed to human G-CSF mRNA, plus a nucleotide sequence taking part in transcriptional control [Benoist and Chambon, Nature, 290, 304 (1981): and Breathnack and Chambon, Ann. Rev. Biochem., 50, 349 (1981)].

(5) Construction of recombinant vector for expression in E. coli (A) +VSE line recombinant vector From the pBRG4 plasmid obtained in (3) (Example 9), a cDNA fragment of the G-CSF polypeptide was cut out with a restriction enzyme arid a recombinant vector was constructed by one of the following methods:
(i) using an annealed synthetic linker, the fragment was ligated with a fragment prepared from a tac promoter-containing pKK223-3 (Pharmacia Fine Chemicals) (Example 12 and Fig. 8);
(ii) three fragments prepared from PL promoter containing pPL-lambda (Pharmacia Fine Chemicals) were ligated with an annealed synthetic linker and, the ligation product and the cDNA fragment were subjected to re-preparation procedures to construct a recombinant vector (Example 13, Fig. 9); or (iii) using an annealed synthetic linker, the fragment was ligated with a fragment prepared from a trp promoter-containing pOYI plasmid (Example 14 and Fig.
10) .
(B) -VSE line recombinant vector In the same manner as described above, three recom-binant vectors were constructed using the plasmid pBRV2 (Example 10) as shown in Example 19 and Figs. 11, 12 and 13.

(6) Preparation of E. coli transformants, and cultivation and expression thereof Using three recombinant vectors of each of the +VSE
and -VSE lines, E, coli strain DH1, N4830 or 3M105 was transformed by the calcium chloride or rubidium chloride procedure described in Molecular Cloning, ibid. (Examples 12, 13, 14 and 19). Each of the transformants obtained was cultivated in ampicillin-containing Luria medium, with induction being subsequently conducted as required to achieve expression (Examples 15 and 20).
'35 (7) Recovery and purification of G-CSF polypeptide from E. coli and amino acid analysis thereof A culture solution of the transformants was centri-f uged to obtain a cell pellet. The collected cells were -ls-treated with a lysozyme and, after lysis by cyclic freezing and thawing, the supernatant was obtained. Alternatively, the cells were treated with guanidium chloride, centrifuged and the supernatant was recovered.
The supernatant was subjected to gel filtration on an Ultrogel ACA54 column (LKB) and the active fractions were concentrated with an ultrafiltration apparatus.
Subsequently, an aqueous solution of trifluoroacetic acid containing n-propanol was added to the concentrate and, after being left in ice, the mixture was centrifuged and adsorbed on a reverse-phase C18 column. After elution, the fractions were checked for their activity. The active frac-tions were collected and subjected to the same procedures of purification as described above. The purified fractions were freeze-dried and the powder was dissolved and subjected to high performance liquid chromatography based on molecular size. The obtained polypeptides were subjected to SDS-polyacrylamide gel electrophoresis and a single band for the desired G-CSF polypeptide was confirmed (Examples 16 and 20). The so obtained polypeptide showed human G-CSF
activity (Examples 17 and 20). The G-CSF polypeptide was analyzed by an amino acid analyzing method with a Hitachi 835 Automatic Amino Acid Analyzer (Hitachi, Ltd.) For analysis of the N-terminal amino acids, a gas-phase sequencer (for Edman decomposition), high-pressure liquid chromatographic apparatus and Ultrasphere-ODS column were used (Examples 18 and 21).
(8) Construction of recombinant vectors for animal cells Recombinant vectors (derived from BPV) for use with C127 and NIH3T3 cells as host cells were constructed for each of the +VSE and -VSE line cDNAs and for the chromosomal gene. Recombinant vectors (with dhfr) for use with CHO
cells were also constructed for each of th +VSE and -VSE
line cDNAs and for the chromosomal gene. Recombinant vectors for use with COS cells were also constructed. In the following, representative examples are described and, for further details, reference should be made to the relevant working examples.

~ 3 41 389 (A) Construction of recombinant vectors of the +VSE line The cDNA (+VSE) fragment obtained in (3) was inserted into a vector pdKCR to make a plasmid pHGA410 (Example 22 and Fig. 14), which was partially digested with EcoRI
followed by treatment with DNA polymerase I (Klenow frag-ment) to create blunt ends. A linker HindIII was attached to the DNA, which was subsequently treated with HindIII
and T4DNA lipase. The treated DNA was used to transform E, coli strain DH1 by the rubidium chloride procedure (see Molecular Cloning, ibid.) The resulting plasmid was named pHGA410(H) (Fig. 15).
The pHGA410(H) was treated with Sall and, after blunt ends were created, it was treated with HindIII once again and a HindIII-SalI fragment was recovered. A plasmid pdBPV-1 having a transformed fragment of bovine papilloma virus was treated with HindIII and PvuII and the larger DNA fragment was separated and joined to the separately prepared HindIII-SaII fragment. The joined fragments were used to transform E. coli strain DH1 to obtain a plasmid, pTN-G4, which had the pHGA410-derived CSF-cDNA (Fig. 15 and Example 23).
Either plasmid, pHGA410 or pHGA410(H), in combination with the plasmid pAdD26SVpA was used to construct pHGG4-dhfr which was a recombinant vector (+VSE) for use with CHO cells (Figs. 16a and b, and Example 25).
A 2-kb DNA fragment containing the dhfr gene was recovered from pAdD26SVpA by treatment with EcoRI and BamHI and the recovered fragment was inserted into pHGA410 (H) at the HindIII site so as to construct pG4DR1 and pG4DR2 (Fig. 16c and Example 25).
(B) Construction of -VSE line recombinant vectors The cDNA (-VSE) fragment obtained in (3) was inserted into a vector pdKCR to make a plasmid pHGV2 (Example 28), which was partially digested with EcoRI followed by treat-ment with DNA polymerase I (Klenow fragment) to create blunt ends. A linker HindIII was attached to the DNA, which was subsequently treated with HindIII and T4DNA
lipase. The treated DNA was used to transform E, coli 1 3 ~~ 389 strain DH1 by the rubidium chloride procedure (see Molecular Cloning, ibid.) The resulting plasmid was named pHGV2(H) (Fig. 18).
The pHGV2(H) was treated with SalI and, after blunt ends were created, it was treated with HindIII once again and a HindIII-SalI fragment was recovered. A plasmid pdBPV-1 having a transformed fragment of vobine papilloma virus was treated with HindIII and PvuII and the larger DNA fragment was separated and joined to the separately prepared HindIII-Sall fragment. The joined fragments were used to transform E. coli strain DHl to obtain a plasmid, pTN-V2, which had the pHGV2-derived CSF-cDNA (Fig. 18 and Example 29).
By similar procedures, either plasmid, pHGV2 or pHGV2(H), in combination with the plasmid pAcID26SVpA was used to construct pHGV2-dhfr which was a recombinant vector (-VSE) for use with CHO cells (Figs. 19a and b, and Example 31).
A DNA fragment of ca. 2 kb containing the dhfr gene 2p was recovered from pAdD,26SVpA by treatment with EcoRI and BamHI and the recover~cl~fragment was inserted into pHGV2 (H) at the HindIII site so as to construct pV2DR1 and pV2DR2 (Fig. 19c and Example 31).
(C) Construction of recombinant vectors containing the chromosomal gene The plasmid gBRCE3~ that was obtained in (4) and which contained the chromosomal gene shown in Fig. 5 was treated with EcoRI.
The pSVH+R+ plasmid described by Banerji et al. in Cell, 27, 299 (1981) was treated with KpnI to remove the globin gene. The plasmid was further subjected to partial digestion with HindIII so as to remove part of the late gene of SV40. The fragments were re-joined to prepare an expression vector pML-E+.
This vector was treated with the restriction enzyme, EcoRI, and dephosphorylated with an alkaline phosphatase (Takara Shuzo Co., Ltd.) to obtain a vector DNA, which was linked to the aforementioned chromosomal DNA fragment with the aid of a T4DNA ligase (Takara Shuzo Co., Ltd.) to obtain pMLCE3a which was a recombinant vector for COS
cells (Example 34). As shown in Fig. 20, this plasmid contained the enhancer of SV40 gene) the replication origin of SV40, the replication origin of pBR322 and the pBR322-derived S-l.ac~.ama:~c~ gene (,~pr ) , and had the human G-CSF chromosomal gene joined downstream from the enhancer of SV4 0 gene .
An expression vector for 0127 cells was constructed by the following procedures. A DNA fragment containing the chromosomal CSF gene was cut out with an appropriate restriction enzyme from pMLCE3a which was the expression vector for COS cells. This fragment was joined, with a T4DNA ligase, to a DNA fragment containing the origin of bovine papilloma virus (BPV) and a DNA fragment containing the early promoter of SV40. The resulting pTNCE3a was an expression vector that had a chromosomal CSF gene linked downstream from the early promoter of SV40 and which contained a 65$ portion of BPV.
The expression vector for CHO cells had two DNA
fragments linked together by a T4DNA ligase; one fragment contained the chromosomal CSF gene and the early promoter of SV40 as in the case of the expression vector for C127 cells, and the other fragment contained a pAdD26SVpA-derived dhfr gene. The resulting pD26SVCE3~ was an expression vector that had the chromosomal CSF gene down-stream of the SV40 promoter and, the dhfr gene downstream of the principal late promoter of adenovirus.
(9) Expression in animal cells Two representative examples are hereunder described and, for further details, see the relevant working examples.
(A) Expression in mouse C127 cells Plasmid pTN-G4 or pTN-V2 was treated with BamHI.
The treated plasmid was used to transform C127 cells (previously grown by cultivation) by the calcium phos-phate procedure. The transformed cells were cultured and clones having high CSF production rate were selected. Glycaproteins containing the expressed G-CSF

1 .341 3gg _22_ were recovered and purified from the culture solution of the transformed cells and were found to have human G-CSF
activity. The presence of the desired glycoprotein was also confirmed by amino acid and sugar content analyses of the sample.
For sugar content analysis, the CSF sample used in amino acid analysis was subjected to determination of amino sugar by the Elson-Margan method, determination of neutral sugar by the orcinol sulfate procedure, or deter-urination of sialic acid by the thiobarbiturate procedure.
The procedures of each determination are shown in "Tohshitsu no Kagaku "Chemistry of Saccharides" (Part 2 of two parts)", Chapter 13, Vol. 4 of A Course in Biochemical Experiments, published by Tokyo Kagaku Dojin. Conversion 1S of the measured values into weight percent revealed that the sugar content of the G-CSF obtained was distributed within the range of 1 - 20 wt$ depending upon the type of host cells, expression vectars and the cultivation conditions.
(B) Expression in COS cells COS cells, which were derived from a monkey CV-1 cells and which had been transformed by SV40-origin deficient mutant to express the large-size T antigen of SV40 [see Gluzman et al., Cell, 32, 175 (1981)], were transformed by the vector pMLCE3a which was obtained in (5)(C) and which contained the human chromosomal G-CSF
gene. The supernatant of the culture of the COS cells showed the human G-CSF activity (Example 35).
The COS cells were recovered and subjected to mRNA
analysis, which showed the existence of two mRNAs that corresponded to the amino acid sequences depicted in Fig.
3(A) and Fig. 4(A), respectively.
Exampl es Before the present invention is described in greater detail with reference to working examples, the following referential examgle is provided for the gurpose of illus-trating the methods of assaying the CSF activity.

~ 3 41 X89 Referential Example: Assaying CSF Activity The following methods were used to determine the CSF
activity (hereunder abbreviated as CSA) in the present invention.
CSA assay (a) With human bone marrow cells: ' Single-layer soft agar cultivation was conducted in accordance with the method of Bradley, T.R. and Metcalf, D.
(Aust. J. Exp. Biol. Med. Sci», 44, 287-300, 1966). More specifically, 0.2 ml of a bovine fetal serum, 0.1 ml of the sample, 0.1 ml of a human bone marrow nonadherent cell sus-pension (1 - 2 x 105 nuclear cells), 0.2 ml of a modified McCoy's 5A culture solutiom, and 0.4 ml of a modified McCoy's 5A culture solution containing 0.?5% of agar were mixed, poured into a plastic dish for tissue culture (35 mm~), coagulated, and cultured at 37°C in 5% C02/95% air and at 100 humidity. Ten days later. the number of colonies formed was counted (one colony consisting of at least 50 cells) and CSA was determined with one unit being the activity required for forming one colony.
(b) With mouse bone marrow cells:
A horse serum (0.4 ml), 0.1 m1 of the sample, 0.1 ml of a C3H/He (female) mouse bone marrow cell suspension (0.5 - 1 x 105 nuclear cells), and 0.4 ml of a modified McCoy's 5A culture solution containing 0.75% of agar were mixed, poured into a plastic dish for tissue culture (35 mm~), coagulated, and cultured fob 5 days at 37°C in 5% C02/95%
air and at 100% humidity. The number of colonies formed was counted (one colony consisting of at least 50 cells) and CSA
was determined with one unit~being the activity fox forming one colony.
The modified McCoy's 5A culture solution used in each of the methods (a) and (b) and the human bone marrow non-adherent cell suspension used in (a) were prepared by the following procedures. ' Modified McCoy's 5A culture solution (double concentration) Twelve grams of McCoy's 5A culture solution (Gibco), 2.55 g of MEM amino acid-vitamin medium (Nissui Seiyaku Co., Ltd.), 2.18 g of sodium bicarbonate and 50,000 units of potassium penicillin G were dissolved twice in 500 ml of distilled water and the solution was aseptically filtered through a Millipore filter (0..22 um).
Human bone marrow nonadherent cell suspension A bone marrow fluid obtained from a healthy person by sternal puncture was diluted 5-fold with an RPMI 1640 culture solution, plated over a Ficoll-Paque solution (Pharmacia Fine Chemicals) and centrifuged at 400 x g for 30 minutes at 25°C. The interfacial cell layer (specific gravity <1.077) was recovered. The cells were washed, adjusted to a concentration of 5 x 106 cells/ml with an RPMI
1640 culture solution containing 20$ of bovine fetal serum, poured into a 25-cm2 plastic flask for tissue culture, and incubated for 30 minutes in a C02 incubator. Nonadherent cells were recovered in the supernatant, poured into a plastic flask (25 cm2) and incubated for 2 hours and a half.
Nonadherent cells in the supernatant were collected and used in an assay.
Example 1: Establishment of CHU-2 A tumor of a patient with oral cancer wherein pro-nounced increase was observed in the number of neutrophiles was transplanted into nu/nu mice. About 10 days after the transplantation, the increase in the weight of the tumor and in the number of neutrophiles was pronounced. Twelve days after the transplantation, the tumor was extracted asepti-cally, shredded into cubes of 1 - 2 mm3 and cultured in the following manner.
Ten to fifteen cubes of the tumor were put into a 50-ml plastic centrifugal tube. After addition of 5 ml of a trypsin solution (containing 0.25 of trypsin and 0.02$
of EDTA), the tube was shaken for 10 minutes in a warm bath at 37°C and the supernatant was discarded. Another 5 ml of the same trypsin solution was added and trypsin digestion was conducted under agitation far 15 minutes at 37°C. The supernatant cell suspension was recovered and stored in ice after the trypsin had been inactivated by addition of 1 ml of a bovine fetal serum.

a After repeating these procedures once again, the cell suspension was recovered, combined with the previously obtained suspension, and centrifuged at 15,000 rpm for 10 minutes to obtain a cell pellet. The pellet was washed twice with F-10 containing 10~ of a bovine fetal serum and was thereafter loaded in a plastic culture flask (25 cm2) to give a cell concentration of 5 x 106 cells/flask. After incubation overnight in a C02 incubator (5~ C02 and 100$
humidity) with an F-10 culture solution containing 10$ of a bovine fetal serum, the supernatant was removed together with the nonadherent cells, and culture was continued with a fresh supply of culture solution. Six days after the start of culture, the flask became full of the cells and the culture solution was replaced by a fresh one. On the next day, the culture solution was discarded and the flask was charged with 2 ml of an anti-mouse erythrocyte antibody (Cappel) diluted 5-fold with RPMI 1640 and 2 ml of a guinea pig complement (Kyokuto Seiyaku Co., Ltd.) diluted 2.5-fold with RPMI 1640. After incubation for 20 minutes at 37°C, the culture was washed twice with F-10 containing 10~ of a bovine fetal serum and the nu~'nu mouse derived fibroblasts were removed. Subsequently, an F-10 culture solution con-taining 10$ of a bovine fetal serum was added and cultiva-tion was conducted for 2 more days. Thereafter, same of the cells were recovered and subjected to cloning by the limiting dilution method.
The resulting 11 clones were checked for their CSF
activity and one clone (CHU-2) exhibited activity about 10 times as high as that of the other clones.
Example 2: Isolation of CSF
The cells established in Example 1 were grown in a completely dense population in two culture flasks (150 cm2).
The cells were recovered, suspended in 500 ml of an F-10 culture solution containing 10$ of a bovine fetal serum, transferred into a glass roller bottle of 1580 cm2 (Belco), and whirl-cultured at 0.5 rpm. When the cells were found to have grown in a completely dense population on the inner wall of the roller bottle, the culture solution was replaced 1 3 ~1 38~

by a serum-free RPMI 1640. After 4-day culture, the super-natant of the culture was recovered and cultivation was continued with F-10 containing 10$ of.a bovine fetal serum being added. After 3-day culture, the culture solution was again replaced by a serum -free RPMI 1640 and the supernatant of the culture was recovered 4 days later. By repeating these procedures, 500 ml of the serum-free supernatant of culture per bottle was obtained each week. In addition, this method enabled the supernatant of culture to be recovered, with the cells maintained over a significantly prolonged period.
A batch consisting of 5,000 ml of the supernatant of the culture obtained was mixed with 0.01 of TWEEN* 20 and concentrated about 1000 times by ultrafiltration with Hollow Fiber DC-4 and Amicon PM-10 (Amicon). The concentrate was purified by the following steps.
(i) A portion (5 ml) of the concentrated supernatant of culture was subjected to gel filtration on an ULT1~AGEL* AcA54 column (4.6 cm~ x 90 cmL~ LKB) at a flow rate of ca. 50 ml/hr with 0.01 M Tris-HC1 buffer (pH 7.4) containing 0.15 M
NaCl and 0.01$ TWEEN 20 (Nakai Kagaku Co., Ltd.) The column had been calibrated with bovine serum albumin (Mw; 67,000), ovoalbumin (Mw: 45,000) and cytochrome C (Mw; 12,400).
After completion of the gel filtration, 0.1 ml of each of the fractions was diluted 10-fold and screened for the active fractions by the above-described method of CSA assay (b). The fractions for Ve = 400 - 700 ml were found to exhibit macrophage-dominant CSA while the fractions for Ve = 800 - 1200 ml showed granulocyte-dominant CSA. There-f ore, the latter fractions were collected and concentrated to a volume of ca. 5 ml on an ultrafiltration apparatus with PM-10 (Amico) .
(ii) To the cocentrated fractions was added an aqueous solution of 0.1~ trifluoroacetic acid coma-fining 30~ of n-. propanol (for determination of amino acid sequence; avail-able from Tokyo Kasei K.K.) After the mixture had been left to stand in ice for about 15 minutes, the precipitate was removed by centrifugation for 10 minutes at 15,000 rpm. The * Trade mark 1 ~ 41 389 supernatant was adsorbed on a u-Bondapak C18 column (8 mm x 30 cm for semipreparatory use; Waters) equilibrated with the aqueous solution containing n-propanol and trifluoroacetic acid; the column was continuously eluted with an aqueous solution of 0.1% trifluoroacetic acid which contained n-propanol having a linear concentration gradient of 30 - 60%.
A high-pressure liquid chromatographic apparatus, Hitachi Model 685-50 (Hitachi, Ltd.), and a detector, Hitachi Model 638-41 (Hitachi, Ltd.) were employed to determine the absprptions at 220 nm and 280 nm simultaneously. After elution, 10 u1 of each of the fractions was diluted 100-fold and screened for the active fractions by the above-described method of CSA assay (b). The peaks eluted with 40% n-propanol were found to have CSA activity, so they were col-lected, re-chromatographed under the same conditions, and assayed for CSA by the same method. Again, CSA activity was observed in the peaks at 40% n-propanol. Therefore, these peaks were collected (4 fractions = 4 ml) and freeze-dried.
(iii) The freeze-dried powder was dissolved in 200 u1 of an aqueous solution of 0.1% trifluoroacetic acid containing 40% of n-propanol, and the solution was subjected to high-pressure liquid chromotography on TSK-G 3000SW column (Toyo Soda Manufacturing Co., Ltd.; 7.5 mm x 60 cm). Elution was conducted with the same aqueous solution at a flow rate of 0.4 ml/min and the fractions were taken in 0.4-ml portions with a fraction collector, FRAC-100 (Pharmacia Fine Chemicals). Each of the fractions taken was checked for its CSA by the same method as described above and activity was observed in the fractions for retention times of 37 - 38 minutes (corresponding to MW of ca. 2 x 104). The active fractions were recovered and purified on an analytical u-Bondapak C18 column (4.6 mm x 30 cm). The main peaks were recovered and freeze-dried. The sample obtained was assayed by the method of CSA assay (a); it was found to have human G-CSF activity.
Examtale 3: Determination of Amino Acid Sequence (i) Determination of N-terminal amino acid sequence The sample was subjected to Edman decomposition with a gas-phase sequencer (Applied Bios;ystems) and the resulting PTH amino acid was analyxed by routine procedures with a high-pressure liquid chromatographic apparatus (Beckman Instruments) and Ultrasphere-ODS column (Beckman Instru-ments). The column {5 um; 4.6 mm~ x 250 mmL) was equi-librated with a starting buffer [aq. sol. containing 15 mM
sodium acetate buffer (pH 4.5 and 40$ acetonitrile] and injected with the sample {as dissolved in 20 u1 of the start-ing buffer). Separation was effected by isocratic elution with the starting buffer. The flow rate was 1.4 ml/min and the column temperature was held at 40°C. Detection of the PTH amino acid was achieved utilizing the absorptions in the UV range at 269 nm and 320 nm. Standard samples of PTH
amino acid (Sigma) in 2-nmol portions were separated on the same line to determine their retention times, which were compare with those of the sample to be tested. As a result, the sample was found to have the following amino acid sequence of the 40 residues from N-terminus:
H2N - Thr - Pro - Leu - Gly - Pro - Ala - Ser - Ser -(10) , ;
Leu - Pro - Gln -~ Ser - Phe - Leu - Leu - Lys - Cys -(20) Leu - Glu - Gln - Val - Arg -- Lys - Ile - Gln - Gly -(30) Asp - Gly - Ala - Ala - Leu - Gln - Glu - Lys - Leu -(40) Cys - Ala - Thr - Tyr - Lys --(ii) Decomposition with bromocyan The sample was dissolved in 70$ formic acid. To the solution, 200 equivalent amounts of bromocyan that had been purified by sublimation was ,added. The mixture was left overnight at 37oC for reaction. The reaction product was freeze-dried and fractionated by HPLC on a TSK G3000SW
column (Toyo Soda Manufacturing Co., Ltd.) to obtain four peaks. The peaks were named CN-1, C:N-2, CN-3 and CN-4 in the decreasing order of the molecular weight. The first two peaks (CN-1 and CN-2) had better yields and their amino acid sequences were analyzed with an automatic gas-phase sequencer (Applied Biosystems) under the same conditions as used in (i).

~ 3 41 389 -29_ As a result, CN-1 was found to be a peptide from the N-terminus of G-CSF protein, and CN-2 had the following amino acid sequence:
Pro - Ala - Phe - Ala - Ser - Ala - Phe Gln - Arg - Arg - Ala - Gly - Gly - Val Leu - Val - Ala - Ser - His - Leu - Gln (iii) Decomposition with trypsin The sample was dissolved in 0.1 M Tris-HC1 buffer (pH

7.4) containing 8 M urea and the solution was mixed with 0.1 M Tris-HC1 buffer (pH 7.4) containing 0.1$ 2-mercaptoethanol to provide a final urea concentration of 2 M. A TPCK
treated trypsin (Sigma) was added such that the sample-to-enzyme ratio was 50:1. The mixture was held for 4 hours at 25oC and, after addition of an equal amount of TPCK-treated trypsin, the mixture was held for an additional 16 hours at 25°C. Thereafter, the reaction product was subjected to high-speed reverse-phased column chromatography on C8 column (Yamamura Kagaku K.K.), with elution conducted with 0.1~ TFA
containing n-propanol having a linear density gradient of 5 l0 - 60$. While several peaks were obtained by measuring the absorption at 280 nm, the main peak was analyzed for its amino acid sequence with an automatic gas-phase sequences (Applied Biosystems) under the same conditions as used in (i). As a result, the main peak was found to be a peptide having the following sequence which contained part of the CN-2 fragment shown in ( i i ) Gln - Leu - Asp - Val - Ala -- Asp - Phe - Ala - Thr -Thr - Ile - Trp - Gln - Gln - Met - Glu - Glu - Leu -Gly - Met - Ala - Pro - Ala - Leu - Gln - Pro - Thr -Gln - Gly - Ala ~- Met - Pro - Ala - Phe - Ala - Ser -Example 4: Preparation of DNA Probe (i) Synthesis of probe (IWQ) Thirty successive nucleotides (see Fig. 1) were prepared on the basis of the sequence of 10 amino acids (Ile-Trp-Gln-Gln-Met-Glu-G1u-Leu-Gly-Met) included within the amino acid sequence obtained in Example 3(iii). It will be necessary to make one comment about the notation of nucleotides shown in Fig. 1; for example, the nucleotide at 9-position from 5'-terminus is an equimolar mixture of dA
and dG. The starting nucleotides were mostly dimers but mononucleotides were also used as required. A glass filter equipped column was charged with 20 mg of the starting nucleotide resin, Ap-d (G) (Yamasa ~r~oyu Co., Ltd. ) After repeated washing with methylene chloride, the 4,4'-dimethoxytrityl group was eliminated by treatment with a solution of methylene chloride containing 3$ trichloroacetic acid. Subsequently, the column was washed several times with 1 ml of methylene chloride. After the column was washed with anhydrous pyridine to displace the solvent, 20 mg of a nucleotide dimer, (DMTr)ApTp(NHR3), (Nippon Zeon;
NHR3 = triethylammonium; DMTr - dimethoxytrityl) and 0.2 ml of pyridine were added, and the interior of the column was vacuum-dried with a vacuum pump. Subsequently, 20 mg of 2,4,6-trimethylbenzenesulfonyl-3-nitrotriazolide (MSNT of Wako Pure Chemical Industries, Ltd.) and 0.2 ml of anhydrous pyridine were added, and the interior of the column was displaced with a nitrogen gas» The nucleotide resin was condensed with the dimer by reaction for 45 minutes at room temperature, with occasional shaking. After completion of the reaction, the column was washed with pyridine and the unreacted OH groups were acetylated with a pyridine solution containing excess acetic anhydride and 4-dimethylamino-pyridine. After washing the column with pyridine, the following dimers or monomers were condensed, in the order written, by repeating the above-described procedures:
(DMTr)Ip(NHR3), (DMTr)GpGp(NHR3), (DMTr)Ip(NHR3), an equimolar mixture of (DMTr)CpTp(NHR3) and (DMTr)TpTp(NHR3), an equimolar mixture of (DMTr)ApAp(NHR3) and (DMTr)ApGp(NHR3), an equimolar mixture of (DMTr)ApGp(NHR~) and (DMTr)GpGp(NHR3), (DMTr)GpAp(NHR3), (DMTr)TpGp(NHR3), an equimolar mixture of (DMTr ) ApAp (NHR3 ) and (DMTr ) GpAp (NHR3 ) , (DMTr)CpAp(NHR3), an equimolar mixture of (DMTr)ApAp(NHR3) and (DMTr)ApGp(NHR3), (DMTr)GpCp(NHR3), (DMTr)TpGp(NHR3), (DMTr)Ip(NHR3) and (DMTr)ApTp(NHR3), with all of these nucleotides being available from Nippon Zeon except for (DMTr)Ip(NHR3) which was available from Yamasa Shoyu Co., Ltd. After completion of the reaction in the final stage, the resin was washed successively with pyridine, methylene chloride and ether without acetylation, and thereafter dried.
The dried resin was suspended in 1.7 ml of a mixture of pyri-dine (0.5 ml), water (0.2 ml) and dioxane (1 ml) containing 1 M tetramethylguanidine and 1 M a-picolinaldoxime. The suspension was left to stand overnight at room temperature and concentrated to 100 - 200 v1 under vacuum. The concen-trate was mixed with a small amount (2 - 3 drops) of pyri-dine and 2 - 3 ml of concentrated aqueous ammonia, and the mixture was heated at 55°G for 6 hours. Following extrac-tion with ethyl aetate, the aqueous layer was separated and concentrated under vacuum. The concentrate was dissolved in a solution of 50 mM triethyl ammonium acetate (pH 7.0) and the solution was subjected to chromato-graphy on C-18 column (1.0 x 15 cm; Waters), with elution conducted with aceto-nitrile (linear density gradient of 10 - 30'x) in a solution of 50 mM triethyl ammonium acetate (pH 7.0). The peak fraction eluted at an acetonitrile concentration of about 25~ was concentrated under vacuum.
To the concentrate, 80~ acetic acid was added and the mixture was left to stand for 30 minutes at room temperature.
Following extraction with ethyl acetate, the aqueous layer was separated and concentrated under vacuum. The resulting concentrate was further purified by high-pressure liquid chromatography on C-18 column (from Senshu Kagaku K.K.;
SSC-ODS-272; 6~ x 200 mm). Elution was conducted with aceto-nitrile (10 - 20~! linear density gradient) in a solution of 50 mM triethyl ammonium acetate (pH 7.0). A synthetic DNA
was obtained in a yield no lower than 10A260 units.
Analysis by the Maxam-Gilbert sequencing method [Meth. Enzym., 65, 499 (1980)) revealed that the oligonucle-otide obtained had the nucleotide sequence shown in Fig. 1.
(ii) Synthesis of probe (A) Fourteen successive nucleotide (see Fig. 1) were obtained on the basis of the sequence of 5 amino acids (Met-Pro-Ala-Phe-Ala) included within the amino acid sequence obtained in Example 3(iii)»

'I 3 4~ 389 Synthesis procedures were similar to those employed in the preparation of probe (IWQ), and the following nucleotides were condensed to a nucleotide resin, Ap-d(T) (Yamasa Shoyu Co., Ltd.) in the order written:
S (DMTr)CpAp(NHR3), (DMTr)GpGp(NHR3), an equimolar mixture of (DMTr)CpAp(NHR3), (DMTr)CpTp(NHR3), (DMTr)CpGp(NHR3) arid (DMTr)CpCp(NHR3), an equimolar mixture of (DMTr)ApGp(NHR3), (DMTr)TpGp(NHR3), (DMTr)GpGp(NHR3) and (DMTr)CpGp(NHR3), (DMTr)ApAp(NHR3), an equimolar mixture of (DMTr)CpAp(NHR3) and (DMTr)CpGp(NHR3), and (DMTr)Gp(NHR3), with all nucle-otides being available from Nippon Zeon. A synthetic DNA
was obtained in a yield of ca, 10A260 units. Analysis by the Maxam-Gilbert sequencing method revealed that the oligo-nucleotide obtained had the nucleotide sequence shown in Fig. 1.
(iii) Synthesis of probe (LC) Automatic DNA synthesis was accomplished with a DNA
synthesizer, Model 380A of Applied Biosystems. This tech-nique, based on the principles described by Caruthers et al.
[J. Am. Chem. Soc., 103, 3185 (1981)], is generally referred to as the ghosphoramidite procedure.
A phosphoramidite form of (DMTr)-dT preliminarily activated with tetrazole was condensed to dG-S (S: support) wherein 5'-dimethoxytrityl group (DMTr) was deblocked.
Thereafter, the unreacted hydroxyl groups were acetylated and oxidated with iodine in the presence of water to make a phosphoryl group. After deblocking the DMTr group, conden-sation was repeated in the same manner until 24 nucleotides having the sequence shown in Fig. 1 were synthesized. These nucleotides were cleaved from the support, deblocked, and purified by reverse-phased high-pressure liquid chromato graphy on C-18 column (Senshu Kagaku Co., Ltd.; SSC-ODS-272).
Example 5: Cultivation of CHU-2 Cells and Preparation of mRNA
1) Cultivation and recovery of CHU-2 cells Established CHU-2 cells were grown in a completely dense population in two culture flasks (150 cm2), recovered, suspended in 500 ml of an RPMI 1640 culture solution ~ 341 3~9 containing 10% of a bovine fetal serum, transferred into a glass roller bottle of 1580 cm2 (Belco), and whirl-cultured for 4 days at 0,5 rpm. When the cells were found to have grown in a completely dense population on the inner wall of the roller bottle, the culture solution was removed from the roller bottle, which was charged with 100 ml of a preheated (37°C) physiological saline solution containing 0.02% of EDTA. After heating at 37°C for 2 minutes, the cells were separated from the inner wall of the flask by pipetting.
The resulting cell suspensian was centrifuged at 1500 rpm for 10 minutes to obtain a cell pellet. The cells were resuspended in 5 ml of an EDTA-free physiological saline.
solution. The suspension was centrifuged at 1500 rpm for 10 minutes to obtain a cell pellet (wet weight, ca. 0.8 g).
The so obtained cells were stared frozen at -80oC until they were subjected to procedures for extraction of RNA.
2) Purification of mRNA
Isolation of mRNA from the CHU-2 cells obtained in 1) was accomplished by procedures which were essentially the same as those described in "Malecular cloning", Maniatis et al., Cold Spring Harbor, page 196, 1982. The frozen CHU-2 cells (wet weight, 3.8 g) were suspended in 20 ml of a solu-tion of 6 M guanidine [6 M guanidinium isothiocyanate, 5 mM
sodium citrate (gH 7,0), 0.1 M ~-mercaptoethanol, and 0.5%
sodium sarcosyl sulfate] and the suspension was well mixed by vortexing for 2 - 3 minutes. The mixture was subjected to 10 cyclic suction and ejection with a syringe (capacity, 20 ml) equipped with a 18G needle. About 6 ml of the viscous guanidinium solution containing the disrupted cells was layered onto a 6-ml cushion of 5.7 M CsCl in 0.1 M EDTA
(pH 7.5) in a Beckman SW40 Ti polyallomer centrifuge tube in such a manner that the tube became full of the contents.
Four centrifuge tubes were prepared by the procedures described above and centrifuged at 30,000 rpm for 15 hours at 20°C. The resulting pellets were washed three times with a small amount of 70% ethanol.
The pellets obtained from the respective tubes were combined, dissolved in 550 u1 of water and worked up to f 1 3 41 389 provide a NaCl concentration of 0.2 M. After treatment with a 1:1 mixture of phenol and chloroform and with chloroform alone, 2.5 volumes of ethanol were added to precipitate the total RNA (ca. 10.1 mg of the total RNA was obtained from 3.8 g of wet cells).
Poly(A+) RNA was purified from the total RNA by the following procedures of affinity chromatography taking advantage of the attachment of a poly(A) chain at 3' termi-nus of the mRNA. Adsorption on oligo(dT)-cellulose (Type 7 of P-L Biochemicals) was achieved by passage through an oligo(dT)-cellulose column of the total RNA in a loading buffer [containing 10 mM Tris-HC1 (pH 7.5), 0.5 M NaCl, 1 mM
EDTA, and 0.1% SDS solution] after the solution had been heated at 65°C for 5 minutes. The column had been equi-librated with the same loading buffer. Elution of poly(A+) RNA was accomplished with a TE solution [containing 10 mM
Tris-HC1 (pH 7.5) and 1 mM EDTA]. The unadsorbed effluent was re-charged through the column and the eluate obtained by repeating the same procedures was mixed with the first run of eluate. As a result, 400 ug of the poly(A+) RNA was obtained.
The so prepared mRNA was fractionated for size by sucrose density gradient centrifugation in accordance with the procedures described in the laboratory manual of Schleif and Wensink, "Practical Methods in Molecular Biology", Springer-Verlag, New York, Heidelberg, Berlin (1981).
Stated more specifically, a 5 - 25% sucrose density gradient was created in a Backman SW40 Ti centrifuge tube.
Two sucrose solutions were prepared by dissolving 5% and 25%
of RNase-free sucrose (Schwarz/Mann) in a solution contain-ing 0.1 M NaCl, 10 mM Tris-HCl (pH 7.5), 1 mM EDTA, and 0.5%
SDS.
Eight hundred micrograms of the mRNA [poly(A+)-RNA]
prepared by the method already described was dissolved in 200 - 500 u1 of a TE solution. The solution was heated at 65°C for 5 minutes and, after being quenched, it was placed on the sucrose density gradient solutions, which were centrifuged at 30,000 rpm for 20 hours. Fractions each X34'389 weighing 0.5 ml were collected and their absorption at 260 nm was measured. The sizes of the fractionated RNAs were determined on the basis of the positions of standard RNAs (ribosome RNAs 285, 18S and 5S). At the same time, the G-CSF activity of each fraction was examined with oocytes of Xenopus laevis by the following procedures, First, the mRNA
of each fraction was worked up into an aqueous solution having a concentration of 1 ~g/ul; oocytes were taken from Xenopus (about one year old) and the mRNA solution was injected in such a manner that a 50-~ng of mRNA was injected into one oocyte; ten such oacytes were placed in each of 96 wells in a microtiter plate; the oocytes were cultured for 48 hours at room temperature in 100 u1 of a Barth medium j88 mM NaCl; 1 mM KC1; 2.4 mM NaHC03; 0.82 mM MgS04; 0.33 mM
Ca(N03)2; 0.41 mM CaCl2; 7.5 mM Tris-HCl (pH 7.6); penicil-lin, 10 mg/L; and streptomycin sulfate, 10 mg/L]; the super-natant of the culture was recovered, concentrated and puri-fied to a grade suitable for assay of G-CSF activity.
The G-CSF activity was found to be present in 15 -17S fractions.
Example 6: Synthesis of cDNA (Construction of pBR-line cDNA Library) From the poly(At) RNA obtained in Example 5 was synthesized cDNA by the method of Land et al. [Nucleic Acids Res., 9, 2251 (1981) ] as modified by the method of Gubler and Hoffman [Gene, 25, 263 (1983)].
1) Synthesis of single-stranded cDNA
An Eppendorf tube (capacity, 1.5 ml) was charged with reagents in the following order: 80 u1 of a reaction buffer (500 mM RC1, 50 mM MgCl2, 250 mM Tris-HC1, pH 8.3); 20 u1 of 200 mM dithiothreitol, 32 u1 of 12.5 mM dNTP (containing 12.5 mM each of dATP, dGTE, dCTP and dTTP), 10 u1 of a-32P-dCTP (PB 10205 of Amerscham) , 32 u1 of oligo (dT) 12-18 (f rom P-L Biochemicals; 500 ug/ml), 20 u1 of poly(A+) RNA (2.1 ug/ul), and 206 u1 of distilled water. A total of 400 u1 of the reaction solution was heated at 65°C for 5 minutes, and thereafter heated at 42°C for 5 minutes. To the heated solution, 120 units of a reverse transcriptase (Takara Shuzo Co., Ltd.) was added. Following reaction for 2 more hours at 42°C, 2 u1 of an RNase inhibitor (Bethesda Research Laboratories), 20 u1 of a TE solution, 16 u1 of 100 mM
sodium pyrophosphate and 48 units (4 ~1) of a reverse tran-scriptase were added, and reaction was carried out at 46°C
f or 2 hours. The reaction was quenched by addition of 0.5 M
EDTA (8 u1) and 10~ SDS (8 ~1). By subsequent treatment with phenol/chloroform and precipitation with ethanol (twice), a single-stranded cDNA was obtained.
2) Attachment of dC-chain to the single-stranded cDNA
The single-stranded cDNA obtained in 1) was dissolved in distilled water. To the solution was added 60 u1 of a dC-chain adding buffer (400 mM potassium cacodylate, 50 mM
Tris-HC1 (pH 6.9), 4 mM dithiothreitol, 1 mM CoCl2, and 1 mM
dCTP], and the mixture was heated at 37°C for 5 minutes. To the reaction solution, 3 p1 of a terminal transferase (27 units/ul; P-L Biochemicals) was added and the mixture was heated at 37oC for 2.5 minutes. Following treatment with phenol/chloroform (once) and precipitation with ethanol (twice), the dC-tailed cDNA was dissolved in 40 u1 of a TE
r solution containing 100 ~n'M NaCI.
3) Synthesis of double-stranded cDNA
To 40 u1 of the DNA solution prepared in 2), 4 u1 of oligo(dG)12-18 (200 ug/ml; P-L Biochemicals) was added and the mixture was heated first at 65oC for 5 minutes, then at 42°C for 30 minutes. While the reaction solution was held at 0°C, 80 u1 of a buffer [100 mM Tris-HCl (pH 7.5), 20 mM
MgCl2, 50 mM (NH4) 2504, and 500 mM TCC1 ] , 4 u1 of 4 mM dNTP
(containing 4 mM each of dATP, dCTP, dGTP and dTTP), 60 u1 of 1 mM S-NAD, 210 u1 of distilled water, 20 u1 of ~ coli DNA polymerase I (Takara Shuzo C:o., Ltd.), 15 u1 of E. coli DNA ligase (Takara Shuzo Co., Ltd.) and 15 u1 of E. coli RNase H (Takara Shuzo Co., Ltd.) were added, and the mixture was subjected to reaction at 12°C for 1 hour. Following addition of 4 mM dNTP (4 u1), reaction was carried out at 25°C for 1 hour. By subsequent treatment with phenol-chloroform and precipitation with ethanol (once), about 8 ~g 1 341 3~9 of a double-stranded cDNA was obtained. This double-stranded cDNA was dissolved in a TE solution and subjected to 1.2~ agarose gel electrophoresis. Fragments correspond-ing to the size of ca. 560 by to 2 kbp were adsorbed on Whatman DE81 and about 0.2 ug of the double-stranded cDNA
could be recovered by elution.
4) Attachment of dC-chain to the double-stranded cDNA
The double-stranded cDNA prepared in 3) was dissolved in 40 u1 of a TE solution. After 8 u1 of a dC-tail adding buffer of the type identified in 2) had been added, the mixture was heated at 37°C for 2 minutes. Following addi-tion of 1 u1 of a terminal transferase (27 units/ul), the mixture was subjected to reaction at 37oC for 3 minutes.
Thereafter, the reaction solution was immediately cooled to 0°C, and the reaction was quenched by addition of 1 u1 of 0.5 M EDTA. Following treatment with phenol/chloroform and precipitation with ethanol, the precipitate obtained was suspended in 10 u1 of a TE solution.
5) Construction of pBR-line cDNA library Four microliters of a commercial oligo(dG)-tailed pBR322 vector (Bethesda Research Laboratories; 10 ng/u1) and 2 u1 of the dC-tailed double-stranded cDNA obtained in 4) were annealed in a TE solution containing 75 u1 of 0.1 M
NaCl. The annealing consisted of three stages: heating at 65°C for 5 minutes, subsequent heating at 40°C for 2 hours, followed by cooling to room temperature.
In accordance with the method described in the laboratory manual of Maniatis et al. [Molecular cloning, Cold Spring Harbor, p 249 ff. (1982)] (other routine tech-niques could also be used here), competent cells were prepared from E. coli strain X1776, and transformed with the annealed plasmid to produce transformants.
Example 7: Synthesis of cDNA (Construction of aphage Library) 1) Synthesis of single-stranded cDNA
In accordance with the procedures described in Example 5, 3.8 g of frozen CHt7-2 cells were purified twice ~1 3 41 389 on an oligo(dT)-cellulose column and subsequently worked up to obtain 400 ug of poly(A+) RNA.
A TE sol ution (10 u1 ) having 12 ug of the poly (A+) RNA dissolved therein was placed in a reaction tube contain-s ing 10 ug of actinomycin D (Sigma). Thereafter, the tube was charged with reagents in the following order: 20 u1 of a reverse transcription buffer [250 mM Tris-HC1 (pH 8.3); 40 mM MgCl2; 250 mM KC1]; 20 u1 of 5 mM dNTP (containing 5 mM
each of dATP, dGTP, dCTP and dTTP); 20 u1 of oligo(dT)12-18 (0.2 ug/ml; P-L Biochemicals); 1 u1 of 1 M dithiothreitol;
2 u1 of RNasin (30 units/ul; Promega Biotech); 10 u1 of a reverse transcriptase (10 units/ul; Seikagaku Rogyo Co., Ltd.); 1 u1 of a-32P-dATP (10 uCi; Amerscham); and 16 u1 of water. The reaction solution totalling a volume of 100 u1 was held at 42°C for 2 hours and the reaction was quenched by addition of 0.5 M EDTA (5 u1) and 20$ SDS (1 u1). By subsequent treatment with phenoi/chlorof orm (100 u1) and precipitation with ethanol (twice), about 4 ug of a single-stranded cDNA was obtained.
2) Synthesis of double-stranded cDNA
The cDNA obtained in 1) was dissolved in 29 u1 of a TE solution and a reaction solution was prepared by adding the following reagents in the order written: 25 u1 of a polymerase buffer [400 mM Hepes (pH 7.6); 16 mM MgCl2, 63 mM
~-mercaptoethanol, and 270 mM KC1]; 10 u1 of 5 mM dNTP; 1.0 u1 Of 15 mM ~-NAD; 1. 0 u1 of a-32-P-dATP (10 uCi/ul ) ; 0. 2 u1 of ~ coli DNA ligase (60 units/ul; Takara Shuzo Co., Ltd.);
5.0 u1 of E. co i DNA polymerase I (New England Biolabs; 10 units/ul); 0.1 u1 of RNase H (60 units/ul; Takara Shuzo Co., Ltd.); and 28.7 u1 of distilled water.
The reaction solution was incubated at 14°C for 1 hour, allowed to return to room temperature, and incubated for an additional hour. Then, the reaction was quenched by addition of 0.5 M EDTA (5 u1) and 20$ SDS (1 u1), and treat-ment with phenol/chloroform and precipitation with ethanol were performed. The DNA obtained was dissolved in 20 u1 of 0.5 mM EDTA and a reaction solution was prepared by addition of 3 u1 of a Rlenow buffer [500 mM Tris-HCl (pH 8.0) and 50 ~ 3 41 389 mM MgCl2], 3 u1 of 5 mM dNTP, and 4 u1 of water. After addition of 1 u1 of a DNA polymerase (Rlenow fragment;
Takara Shuzo Co., Ltd.), the reaction solution was incubated at 30oC for 15 minutes.
The incubated reaction solution was diluted with 70 u1 of a TE solution and the reaction was quenched by addi-tion of 0.5 M EDTA (5 u1) and 20% SDS (1 u1). By subsequent treatment with phenol/chloroform and precipitation with ethanol, about 8 up of a double-stranded cDNA was obtained.
3) Methylation of double-stranded cDNA
An aqueous solution (30 u1) of the double-stranded cDNA synthesized in 2) was mixed with 40 u1 of a methylation buffer [500 mM Tris-HC1 (pH 8.0); 50 mM EDTA], 20 u1 of a SAM solution [800 uM S-adenosyl-L-methylmethionine (SAM); 50 mM ~-mercaptoethanol], and 100 u1 of water. To the mixture, 15 j~l of an EcoRI methylase (New England Biolabs; 20 units/ul) was added to make a reaction solution totalling 200 u1 in volume. Following incubation at 37°C for 2 hours, treatments with phenol and ether and precipitation with ethanol were conducted to recover the DNA.
4) Addition of EcoRI linker To about 1.2 up of the methylated double-stranded DNA, 1.5 u1 of a lipase buffer [250 mM Tris-HC1 (pH 7.5) and 100 mM MgCl2], 0.5 u1 of a preliminarily phosphorylated EcoRI linker (lOmer; Takara Shuzo Co., Ltd.), 1.5 u1 of 10 mM ATP, 1.5 u1 of 100 mM dithiothreitol, and 2 u1 of H20 were added to make a reaction solution totalling 15 u1 in volume. After 0.7 u1 of T4 DNA lipase (3.4 units/ul; Takara Shuzo Co., Ltd.) had been added, reaction was carried out overnight at 4°C. Thereafter, the l.igase was inactivated by heating at 65°C for 10 minutes. The reaction solution was worked up to a total volume of 50 ~.~1 by addition of 100 mM
Tris-HC1 (pH 7.5), 5 mM MgCl2, 50 mM NaCl and 100 ug/ml of gelatin. Following addition of EcoRI (3.5 u1: 10 units/ul), reaction was carried out at 37°C for 2 hours. Subsequently, 2.5 u1 of 0.5 M EDTA and 0.5 u1 of 20% SDS were added, followed by treatment with phenol/chloroform and precipita-tion with ethanol so as to recover the DNA. Thereafter, the 1 34~ 389 unreacted EcoRI linker was removed by gel filtration on Ultrogel AcA34 (LKB) or agarose-gel electrophoresis, so as to recover about 0.5 - 0.7 1~g of the linker-added double-stranded cDNA.
5) Joining double-stranded cDNA to agtl0 vector The linker-added double-stranded cDNA was mixed with 2.4 up of preliminarily EcoRI-treated ~,gtl0 vector (Vector Cloning system), 1.4 u1 of a lipase buffer (250 mM Tris-HC1 and 100 mM MgCl2), and 6.5 u1 of distilled water, and the mixture was heated at 42°G for 15 minutes. Thereafter, 1 u1 of 10 mM ATP, 1 u1 of 0.1 M dithiothreitol and 0.5 u1 of T4 DNA lipase were added to make a total volume of 15 u1 and reaction was carried out overnight at 12°C.
6) In vitro packaging About a third of the recombinant DNAs prepared in 5) was packed with an in vitro packaging kit (Promega Biotech) to obtain phage plaques.
Example 8: Screening of pBR-Line Library with Probe (IWQ) Whatman 541 paper was placed on a colony-growing agar .medium and left to stand at 37oC for 2 hours. The filter paper was subsequently tfeated by the following method of Taub and Thompson [Anal. Biochem., x26, 222 (1982)].
The colonies transferred onto the 541 paper was further grown onto an agar medium containing chloramphenicol (250 ug/ul) overnight at 37°C.
The 541 paper was recovered and left at room tempera-ture for 3 minutes on another sheet of filter paper that had been impregnated with a 0.5 N NaOH solution. This procedure was repeated twice. Two similar runs were conducted for 3 minutes using a solution of D.5 M Tris-HCl (pH 8). At 4°C, treatments were conducted with a solution of 0.05 M Tris-HC1 (pH 8) for 3 minutes, and with 1.5 mg/ml of a lysozyme solution (containing 0.05 M Tris-HC1 (pH 8) and 25$ sucrose]
for 10 minutes; then, at 37°C, treatments were conducted with a solution of 1 x SSC (0.15 M NaCl and 0.015 M sodium citrate) for 2 minutes, and with a 1 x SSC solution contain-ing 200 ug/ml of proteinase K for 30 minutes; finally, at room temperature, treatments were conducted with a 1 x SSC

solution for 2 minutes, and with 95% ethanol solution for 2 minutes. The final step was repeated twice. Thereafter, the 541 paper was dried. The dried 541 paper was immersed in a 25:24:1 mixture of phenol/chlaroform/isoamylalcohol [equilibrated with 100 mM Tris-HC1 (pH 8.5), 100 mM NaCl and 10 mM EDTA] for 30 minutes at room temperature. Subse-quently, similar procedures were repeated three times with a 5 x SSC solution for 3 minutes, then twice with a 95%
ethanol solution for 3 minutes. Thereafter, the filter paper was dried.
The probe (IWQ) was labelled with 32P in accordance with routine procedures (see Molecular Cloning) and colony hybridization was performed in accordance with the method of Wallace et al. [Nucleic Acids Res., 9, 879 (1981)].
Prehybridization was conducted at 65°C for 4 hours in a hybridization buffer containing 6 x NET [0.9 M NaCl; 0.09 M
Tris-HC1 (pH 7.5): and 6 mM EDTA], 5 x Denhardt's solution, 0.1% SDS and 0.1 mg/ml of denatured DNA (calf thymus).
Thereafter, hybridization was conducted overnight at 56°C in a hybridization buffer (f or its formulation, see above) containing 1 x 106 cpm/ml of the radiolabelled probe (IWQ).
After completion of the reaction, the 541 paper was washed twice with a 6 x SSC solution (containing 0.1% SDS) f or 30 minutes at room temperature, then washed at 56°C for 1.5 minutes. The washed 541 paper was then subjected to autoradiography.
The plasmid was separated from positive clones and subjected to Southern blotting with the probe (IWQ).
Hybridization and autoradiography were conducted under the same conditions as described above.
Similarly, Southern blotting was conducted with the probe (A). Using a hybridization buffer having the formula-tion shown above, hybridization was conducted first at 49°C
for 1 hour. After leaving it to 39°C, hybridization was further continued at the same temperature for 1 hour. After completion of the reaction, a nitrocellulose filter was washed twice With 6 x CSC cont:aini.n~ 0.1~ SDS for 30 minutes a 1 34~ 3~9 at room temperature, then washed at 39°C for 3 minutes. The washed paper was subjected to autoradiography.
As a result, a single clone was found to be positive,.
Nucleotide sequencing by the dideoxy method revealed that this clone had a DNA composed of 308 base pairs containing the portions of both probe (IWQ) and probe (A). The pBR322-derived plasmid containing this insert was named pHCS-1.
Example 9: Screening of lPhage Line Library with pHCS-1 Derived DNA Probe Plaque hybridization was conducted in accordance with the method of Benton and Davis [Science, 196, 180 (1977)].
The pHCS-1 obtained in Example 8 was treated with Sau3A and EcoRI to obtain a DNA fragment of ca. 600 bp. This DNA
fragment was radiolabelled by nick translation in accordance with routine procedures. A nitrocellulose filter (S & S) was placed on the phage plaque-growing agar medium to trans-fer the phages onto the filter. After denaturing the phage DNA with 0.5 M NaOH, the filter paper was treated by the following procedures: treatment with 0.1 M NaOH and 1.5 M
NaCl for 20 seconds; two treatments with 0.5 M Tris-HC1 (pH
7.5) and 1.5 M NaCl for 20 seconds; finally, treatment with 120 mM NaCl, 15 mM sodium citrate, 13 mM KH2P04 and 1 mM
EDTA (pH 7.2) for 20 seconds.
The filter was subsequently dried and heated at 80°C
for 2 hours to immobilize the DNA. Prehybridization was conducted overnight at 42oC in a prehybridization buffer containing 5 x SSC, 5 x Denhardt's solution, 50 mM phosphate buffer, 50% formamide, 0.25 mg/ml of denatured DNA (salmon sperm DNA) and 0.1% SDS. Thereafter, hybridization was conducted at 42°C for 20 hours in a hybridization buffer containing 4 x 105 cpm/ml of pHCS-1 probe that had been radiolabelled by nick translation. This hybridization buffer was a mixture of 5 x SSC, 5 x Denhardt's solution, 20 mM phosphate buffer (pH 6.0), 50% formamide, 0.1% SDS, 10% dextran sulfate and 0.1 mg/ml of denatured DNA (salmon sperm DNA).

'i ~ 41 389 The hybridized nitrocellulose filter was washed for 20 minutes with 2 x SSC containing 0,1% SDS at room tempera-ture, then for 30 minutes with 0.1 x SSC containing 0.1% SDS
at 44oC, and finally for 10 minutes with 0.1 x SSC at room temperature. Detection by autoradiography was then conducted.
As a result, five positive clones (G1 - G5) were obtained. The clone contained a "full-length" cDNA was checked for its DNA nucleotide sequence by the dideoxy method and the nucleotide sequence shown in Fig. 3(A) was identified. This cDNA was cut out of the ~gtl0 vector and joined to pBR327 [Soberon et al., Gene, _9, 287 (1980)] at the EcoRI site to form a plasmid which could be prepared on a large scale. This plasmid i.s named pBRG4.
Example 10: Screening of ~Phage Line Library with pBRG4-Derived DNA Probe and Probe (LC) Plaque hybridization was performed in accordance with the method of Benton and Davis (see Science, ibid.) employed in Example 9. A nitrocellulose filter (S & S) was placed on the phage plaque-growing agar medium to transfer the phages onto the filter. After denaturing the phage DNA with 0.5 M
NaOH, the filter was treated by the following procedures:
treatment with 0.1 M NaOH and 1.5 M NaCl for 20 seconds;
then two treatments with 0.5 M Tris--HC1 (pH 7.5) and 1.5 M
NaCl for 20 seconds; finally, treatment with 120 mM NaCl, 15 mM sodium citrate, 13 mM KH2P04 and 1 mM EDTA (pH 7.2) for 20 seconds. The filter was subsequently dried, and heated at 80oC for 2 hours to immobilize tree DNA. Two sheets of the same filter were prepared in the manner described above and subjected to screening with the pBRG4-derived DNA probe and the probe (LC).
Screening with the pBRG4-derived DNA probe was carried out by the following pracedures. The pBRG4 was treated with EcoRI to obtain a DNA fragment of ca. 1500 bp.
This DNA fragment was radiolabelled by nick translation in accordance with routine procedures. One of the two nitro-cellulose filters was subjected to prehybridization over-night at 42°C in a prehybridization buffer containing 5 x SSC, 5 x Denhardt's solution, 50 mM phosphate buffer, 50%

~ 3 41 389 _44_ formamide, 0.25 mg/ml of denatured DNA (salmon sperm DNA) and 0.1% SDS. Thereafter, the filter was subjected to hybridization at 42°C for 20 hours in a hybridization buffer containing the radiolabelled DNA probe (ca. 1 x 106 cpm/ml) of ca. 1500 bp. This hybridization buffer was a mixture of 5 x SSC, 5 x Denhardt's solution, 20 mM phosphate buffer (pH
6.0), 50% formamide, 0.1% SDS, 10% dextran sulfate and 0.1 mg/ml of denatured DNA (salmon sperm DNA). The hybridized nitrocellulose filter was washed for 20 minutes with 2 x SSC
containing 0.1% SDS at room temperature, then for 30 minutes with 0.1 x SSC containing 0.1% SDS at 44oC, and finally for 10 minutes with O.lx SSC at room temperature. Detection by auto.radiography was then conducted.
Screening with the probe (LC) was carried out by the following procedures. The other filter was preliminarily treated with 3 x SSC containing 0.1% SDS at 65°C for 2 hours.
Then, prehybridization was conducted at 65°C for 2 hours in a solution containing 6 x NET, 1 x Denhardt's solution, and 100 ug/ml of denatured DNA (salmon sperm DNA). Hybridiza-tion was subsequently conducted overnight at 63°C in a hybridization buffer containing the radiolabelled probe (LC) (2 x 106 cpm/ml). This hybridization buffer was also a mixture of 6 x NET, 1 x Denhardt's solution and 100 ug/ml of denatured DNA (salmon sperm DNA). The hybridized nitro-cellulose filter was washed three times (20 minutes each) with 6 x SSC containing 0.1% SDS at room temperature, then washed with 6 x SSC containing 0.1% SDS at 63°C for 2 minutes.
The filter was dried and detection was conducted by autoradiography.
In the screening described above, clones which were positive to both probes we-re selected and the clone con-tained a "full-length" cDNA was checked for its nucleotide sequence by the dideoxy method. It was found to have the nucleotide sequence shown in Fig. 4(A). This cDNA was cut out of the ~gtl0 vector and joined to pBR327 at the EcoRI
site to prepare a plasmid pBRV2.
Example 11: Screening of Human Chromosomal Gene Library 1 34~ 389 _45-1) Construction of human chromosomal gene library The human chromosomal gene library which was provided by courtesy of Dr. Maniatis of Harvard University had been prepared by the following procedures: the whole chromosomal DNA was extracted from the human fetal liver with phenol or other appropriate chemicals and partially digested with restriction enzymes, HaeITI and Alul; the resulting DNA
fragments were treated by sucrose density gradient centrif-ugation to concentrate the fragments having chain lengths of about 18 - 25 kb; the concentrated fragments were joined to the arm DNA of E. coli phage ~ Charon 4A, with short-chained synthetic nucleotides having the cleavage sites of the restriction enzyme EcoRI being inserted, so as to prepare infectious phage DNA recombinants; with a view to providing enhanced infectiousness, more refined phage a particles were created by packaging. The so prepared human gene library is theoretically considered to be a set of recombinants con-taining human DNAs with chain lengths of 18 - 25 kb which contained practically all human genes.
2) Screening of human chromosomal gene library with the pH CS-1 derived DNA pX'obe Plaque hybridization was conducted in accordance with the method of Benton and Davis [Science, 196, 180 (1977)].
The pHCS-1 obtained in Example 8 was treated with Sau3A and EcoRI to obtain a DNA fragment of ca. 600 bp. This DNA
fragment was radiolabelled by nick translation in accordance with routine procedures. A'nitrocellulose filter (S & S) was placed on the phage plaque-growing agar medium to trans-fer the phages onto the filter. After denaturing the phage DNA with 0.5 M NaOH, the filter paper was treated by the following procedures: treatment with 0.1 M NaOH and 1.5 M
NaCl for 20 seconds; two treatments with 0.5 M Tris-HC1 (pH
7.5) and 1.5 M Nac:l for 20 seconds; finally, treatment with 120 mM NaCl, 15 mM sodium citrate, 13 mM KH2P04 and 1 mM
EDTA (pH 7.2) for 20 seconds.
The filter was subsequently dried and heated at 80°C
for 2 hours to immobilize the DNA. Prehybridization was conducted overnight at 42°C in a prehybridization buffer containing 5 x SSG, 5 x Denhardt's solution, 50 mM phosphate buffer, 50% formamide, 0.25 mg/ml of denatured DNA (salmon sperm DNA) and 0.1% SDS. Thereafter, hybridization was conducted at 42°C for 20 hours in a hybridization buffer containing 4 x 105 cpm/ml of pHCS-1 probe that had been radiolabelled by nick translation. This hybridization buffer was a mixture of 5 x SSC, 5 x Denhardt's solution, 20 mM phosphate buffer (pH 6.0), 50% formamide, 0.1% SDS, 10%
dextran sulfate and 0.1 mg/ml of denatured DNA (salmon sperm DNA).
The hybridized nitrocellulose filter was washed for minutes with 2 x SSC containing 0.1~ SDS at room tempera-ture, then for 30 minutes with 0.1 x SSC containing 0.1% SDS
at 44°C, and finally for 10 minutes with 0.1 x SSC at room 15 temperature. Detection by autoradiography was then conducted.
As a result, ten-odd positive clones were obtained.
Recombinant DNAs were prepared from these clones by the method of Maniatis [Cell, 15, 687 (1978)]. The obtained 20 DNAs were treated with restriction enzymes such as EcoRI, BamHI and BglI2, analyzed by agarose gel electrophoresis, and their restriction enzyme map was prepared in accordance with the method of Fritsch et al. (see cell, ibid.) Southern hybridization was conducted with the probe being the radiolabelled pHCS-1 derived DNA fragment that was the same as what was used in the above-described screening procedures. A DNA fragment of ca. 8 kbp that was cut with EcoRI was selected from the clones that hybridized with the probe. This fragment was subcloned to the EcoRI site of pBR327. The subcloned DNA was subjected to another treat-ment with restriction enzymes and Southern hybridization was conducted repeatedly. A DNA fragment of ca. 4 kbp that was cut out with EcoRI and XhoI was found to contain a gene coding for the human G-CSF polypeptide. This DNA fragment was checked for the sequence of its ca. 3-kbp portion by the dideoxy method and the nucleotide sequence shown in Fig. 5 was identified» This DNA fragment had the restriction enzyme cleavage sites shown in Fig. 7.

Screening of human chromosomal genes was also conducted using pBRG4-derived DNA and pBRV2-derived DNA as probes. In either case, a DNA fragment of 1500 by that had been treated with EcoRI was directly radiolabelled by nick translation in the manner described above or, alternatively, a DNA fragment of ca. 700 by that was obtained by successive treatments with EcoRI and Drar was radiolabell.ed by nick translation. The so prepared probe was used in plaque hybridization that was conducted under the same conditions as described above. Selected clones were analyzed by Southern hybridization so as to obtain a DNA fragment having the nucleotide sequence shown in Fig. 5. The plasmid thus obtained was named pBRCE3,,~:.
Example 12: Construction of E. coli Recombinant Vector (+VSE) and Transformation (Using tac Promoter-Containing Vector) 1) Construction of recombinant vector (i) Vector preparation Five micrograms of a tac promoter-containing vector pKK223-3 (Pharmacia) was treated with 8 units of EcoRI
(Takara Shuzo Co., Ltd.) for 2 hours at 37°C in 30 u1 of a reaction solution (40 mM Tris-HC1, 7 mM MgCl2, 100 mM
NaCl, and 7 mM 2-mercaptoethanol).
Subsequently, 3 u1 of an alkali phosphatase (Takara Shuzo Co., Ltd.) was added and treatment was conducted at 60°C for 30 minutes. A DNA fragment was recovered by three treatments with phenol, one treatment with ether and precipitation with ethanol, all being conducted in accordance with routine procedures.
The recovered DNA fragment was dissolved in a 50-ul mixture composed of 50 mM Tris-HC1, 5 mM MgCl2, 10 mM
DTT, and 1 mM each of dATP, dCTP, dGTP and dTTP. After addition of 3 u1 of an E. co i DNA polymerase I - Klenow fragment (Takara Shuzo Co., Ltd.), reaction was carried out at 14°C for 2 hours to create blunt ends.
(ii) Preparation of synthetic linker Three micrograms of oligonucleotides having the sequences of synthetic linkers, CGAATGACCCCCCTGGGCC and CAGGGGGGTCATTCG, was phosphorylated by performing reac-tion in 40 u1 of a reaction solution (composed of 50 mM
Tris-HC1, 10 mM MgCl2, 10 mM 2-mercaptoethanol and 1 mM
ATP) at 37°C for 60 minutes in the presence of 4 units of T4 polynucleotide kinase.
Each of the phosphorylated oligonucleotides (0.2 up) was dissolved in 20 u1 of a 100 mM NaCl-containing TE
solution [10 mM Tris-HC1 (pH 8.0) and 1 mM EDTA]. After treatment at 65°C for 10 minutes, the oligonucleotides were annealed by slow cooling to room temperature.
(iii) Preparation of G-CSF cDNA fragment Sixty micrograms of the pBRG4 prepared in Example 9 which contained the cDNA shown in Fig. 3(A) was treated with 100 units of a restriction enzyme ApaI (New England Biolabs) and 50 units of DraI (Takara Shuzo Co., Ltd.) at 37°C for 3 hours in 200 1~1 of a reaction solution composed of 6 mM Tris-HC1, 6 mM MgCl2, and 6 mM 2-mercaptoethanol. About 2 up of an ApaI - DraI fragment (ca. 590 bp) was recovered by 1.2~ agarose gel electrophoresis.
(iv) Ligation of fragments About 0.1 up each of the fragments prepared in (i) to (iii) was dissolved in 20 u1 of a ligation solution (66 mM Tris-HC1, 6.6 mM MgCl2, 10 mM DTT, and 1 mM ATP).
After addition of 175 units of T4 DNA lipase, the solu-tion was held overnight at 4oC to obtain a recombinant vector (Fig. 8).
2) Transformation Using 20 u1 of a reaction solution containing the recombinant vector prepared in (iv), E, coli strain JM105 was transformed by the rubidium chloride procedure [see T. Maniatis et al., Molecular Cloning, p. 252 (1982)]. The plasmid was separated from an ampicillin-resistant colony culture of the transformants and treated with restriction enzymes, BamHI, AccII and ApaI to confirm that the transfor-mants were the intended ones.

1 34~ 389 -4 g--Example 13: Constructian of E. co i Recombinant Vector (+VSE) and Transformation (Using PL Promoter-Containing Vector) 1) Construction of recombinant vector (i) Vector preparation A hundred micrograms of a PL promoter-containing vector pPL-lambda (Pharmacia) was treated overnight at 37°C with 50 units of a restriction enzyme BamHI in 100 u1 of a reaction solution [10 mM Tris-HCl (pH 7.6), 7 mM
MgCl2, 100 mM NaCl, and 10 mM DTTj.
By subjecting the reaction solution to 1~ agarose gel electrophoresis, about 49 ug of an approximately 4-kb fragment and about 11 ug of an approximately 1.2-kb fragment were recovered.
The 4-kb fragment was dissolved in 100 u1 of a TE
buffer (for its composition, see above) and dephosphory-lated by reaction with an alkali phosphatase (Takara Shuzo Co., Ltd.) at 60°C for 60 minutes.
The other fragment of about 1.2 kb in length was dissolved in 20 u1 of a buffer (10 mM Tris-HC1, 10 mM
MgCl2, 6 mM KC1, and 1 mM DTT) and treated overnight with 20 units of a restriction enzyme MboII (New England Biolabs) at 37°C.
By 4$ polyacrylamide gel electrophoresis, about 0.9 u9 of a BamHI-MboII fragment (ca. 200 bp) and about 1.9 ug of an MboII-BamHI fragment (ca. 310 bp) were recovered.
(ii) Preparation of synthetic linker Oligonucleotides having the sequences of synthetic linkers, TAAGGAGAATTCATCGAT and TCGATGAATTCTCCTTAG, were phosphorylated and annealed as in (ii) in Example 12, so as to prepare a synthetic S/D linker.
(iii) Preparation of expression vector One tenth of a microgram of the ca. 4-kb fragment, 0.05 ug each of the BamHI-MboII fragment having the OLPL
region and the MboII-BamHI fragment having the tLl region [the three fragments being prepared in (i)j, and 0.1 u9 of the annealed synthetic S/D linker prepared in ~ 341 X89 (ii) were subjected to reaction overnight at 12°C in 40 u1 of a reaction solution (66 mM Tris-HC1, 6.6 mM MgCl2, mM DTT, and 1 mM ATP) in the presence of 175 units of T4 DNA ligase (Takara Shuzo Co., Ltd.) Twenty micro-s liters of the reaction solution was used to transform E.
coli strain N99CI+ (Pharmacia) by the calcium chloride procedure (see Molecular Cloning, ibid.) The transformants were cultured and the plasmid was recovered from the culture of their ampicillin-resistant 10 colonies. Treatment of the plasmid with restriction enzymes, EcoRI, BamHI and Smal, showed that it was the intended plasmid.
Two micrograms of-this plasmid was reacted with a restriction enzyme ClaI (New England Biolabs) at 37°C
for 2 hours in 20 u1 of a buffer (10 mM Tris-HC1, 6 mM
MgCl2 and 50 mM NaCl). Thereafter, the enzyme was inactivated by heating at 65°C for 10 minutes.
One microliter of the reaction solution was reacted overnight at 12°C with 175 units of T4 DNA ligase (Takara Shuzo Co., Ltd.) in a ligation solution having the composition descfibed above. The reaction solution was then used to transform E~, coli strain N99cI+
(Pharmacia). The plasmid was recovered from the culture of ampicillin-resistant colonies of the transformants and treated with EcoRI and BamHI to confirm that said plasmid was the intended one.
(iv) Preparation of G-CSF expressing recombinant vector and transf ormants The expression plasmid prepared in (iii) was treated with a restrict ion enzyme ClaI. After creating blunt ends,_the plasmid was then worked up as in Example 12 to prepare a recombinant vector inserted a cDNA
fragment of G-CSF. This vector was used to transform E.
coli strain N4830 (Pharmacies Fine Chemicals) by the calcium chloride procedure described in Molecular Cloning (ibid. ) Identification of the desired transfor-mants was achieved as in Example 12 (Fig. 9).

'~ 3 41 3~9 Example 14: Construction of ~. coli~~ecombinant Vector (+VSE) and Transformation (Using trp Promoter-Containing Vector) 1) Construction of recombinant vector (i) Vector preparation A plasmid, p0Yl, was prepared by inserting a tryptophan promoter containing HpaII-TaqI fragment (ca.
330 bp) into pBR322 at the CIaI site. Ten micrograms of this plasmid was treated with 7 units of a restriction enzyme ClaI and 8 units of PvuII at 37°C for 3 hours in 30 u1 of a reaction solution composed of 10 mM Tris-HC1, 6 mM MgCl2 and 50 mM NaCl.
Subsequently, 2 ~,l of an alkali phosphatase (Takara Shuzo Co., Ltd.) was added and reaction was carried out at 60°C for 1 hour.
A DNA fragment (ca. 2.5 ug) of about 2.6 kb in length was recovered from the reaction solution by 1%
agarose gel electrophoresis.
(ii) Preparation of Synthetic linker Oligonucleotides having the sequences of synthetic linkers, CGCGAATGACCCCCCTGGGCC and CAGGGGGGTCATTCG, were phosphorylated and annealed as in (ii) in Example 12, so as to prepare a synthetic linker.
(iii) Preparation of recombinant vector About 1 ug of the vector fragment prepared in (i), about 1 ug of the synthetic linker prepared in (ii) and about 1 ug of the G-CSF cDNA fragment prepared in (iii) in Example 12 were reacted with 175 units of T4 DNA
ligase overnight at 12°C in 20 u1 of a ligation solution having the formulation described in Example 12, 1)(iv), so as to obtain a recombinant vector (Fig. 10).
2) Transformation Twenty microliters of the reaction solution prepared in (iii) was used to transform ,E~ co i DHl by the rubidium chloride procedure described in Molecular Cloning, ibid.
As in Example 12, the plasmid was recovered from amplicillin-resistant colonies of the transformants, and treatment of this plasmid with restriction enzymes, Apal, 1 34~ 389 DraI, NruI and Pstl, showed that the desired transformants had been obtained.
Example 15: Cultivation of Transformants 1) Cultivation of the transformants (with tac) obtained in Example 12 The transformants were cultured overnight at 37°C, and 1 ml of the culture was added to 100 ml of a Luria medium containing 25 ug/ml ar 50 ug/ml of amplicillin.
Cultivation was conducted for 2 - 3 hours at 37°C.
The cultivation was continued at 37°C for 2 - 4 hours after addition of isopropyl--8-D-thiogalactoside to make final concentration to 2 mM.
2) Cultivation of the transformants (with PL) obtained . in Example 13 The transformants were cultured overnight at 28oC, and 1 ml of the culture was added to 100 ml of a Luria medium containing 25 or 50 ~g/ml of ampicillin. Cultivation was conducted for about 4 hours at 28°C.
The cultivation was continued for 2 - 4 hours at 42°C.
3) Cultivation of the transfarmants (with trp) obtained in Example 14 The transformants were cultured overnight at 37°C, and 1 ml of the culture was added to 100 ml of M9 medium containing 0.5$ glucose, 0.5$ Casamino acids (Difco) and 25 or 50 ug/ml of ampicillin. Cultivation was conducted for 4 - 6 hours at 37°C. After addition of 50 ug/ml of 3-S-indolacrylic acid (IAA), the cultivation was continued for 4 - 8 hours at 37°C.
Example 16: Recovery and Purification of G-CSF Polypeptide from E: coli 1) Recovery The three species of transformants cultured in Example 15 were subjected to the fallowing recovery procedures.
The culture (100 ml) was centrifuged to obtain a cell pellet, which was suspended in 5 ml of a mixture of 20 mM
Tris-HC1 (pH 7.5) and 30 mM NaCI.

r Then, 0.2 M phenylmethylsulfonyl fluoride, 0.2 M EDTA
and a lysozyme were added in respective concentrations of 1 mM, 10 mM and 0.2 mg/ml, and the suspension was left for 30 minutes at 0°C.
The cells were lyzed by three cycles of freezing/
thawing, followed by optianal sonication. The lysate was centrifuged to obtain the supernatant. Alternatively, the lysate was treated with 8 M guanidine hydrochloride such that its final concentration was 6 M guanidine hydro-chloride, followed by centrifugation at 30,000 rpm for 5 hours, and recovery of the supernatant.
2) Purification (i) The supernatant obtained in 1) was subjected to gel filtration on an Ultrogel AcA54 column (4.6 cm~ x 90 cmL; LKB) at a flow rate of ca. 50 ml/hr with 0.01 M
Tris-HC1 buffer (pH 7.4) containing 0.15 M NaCl and 0.01% Tween 20 (Nakai Ragaku Co., Ltd.) The fractions which showed activity upon analysis by the method of CSA assay (b) (described earlier in this specification) were selected and concentrated to a volume of ca. 5 ml with an ultrafiltration apparatus, pM-10 (Amicon) .
(ii) To the concentrated fractions were added n-propanol (of the grade suitable for amino acid sequencing; Tokyo Kasei Co., Ltd.) and trifluoroacetic acid, and the mix ture was worked up such that the final concentrations of n-propanol and trifluoroacetic acid were 30% and 0.1%, respectively. The worked up mixture was left in ice for about 15 minutes and centrifuged at 15,000 rpm for 10 minutes to remove the precipitate. The supernatant was adsorbed on a u-Bondapak C18 column (of semipreparatory grade: Waters; 8 mm x 30 cm) that had been equilibrated with an aqueous solution containing n-proganol (see above) and trifluoroacetic acid. The column was contin-uously eluted with an aqueous solution of 0.1% tri-fluoroacetic acid containing n-propanol with a linear density gradient of 30 - 60%. ~fith Hitachi Model 685-50 (high-pressure liquid chromatographic apparatus of Hitachi, Ltd.) and Hitachi Model 638-41 (detector of Hitachi, Ltd.) being used, the adsorptions at 220 nm and 280 nm were measured simultaneously. After eluting, a 10-ul aliquot of each fraction was diluted 100-fold and the dilutions were screened for active fractions by the method of CSA assay (b). Activity was observed in the peaks that were eluted at 40% n-propanol. These peaks were combined and re-chromatographed under the same conditions as used above and the fractions were checked for their activity by the method (b). Again, activity was found in the peaks for 40% n-propanol. These active peaks were collected (four fractions = 4 ml) and freeze-dried.
(iii) The freeze-dried powder was dissolved in 200 u1 of an aqueous solution of 0.1% trifluoroacetic acid con-taining 40% n-propanol, and the solution was subjected to high-pressure liquid chromatography on TSK-G3000SW
column (7.5 mm x 60 cm: Toyo Soda Manufacturing Co., Ltd.) Elution was conducted at a flow rate of 0.4 2p ml/min with an aqueous solution of 0.1$ trifluoroacetic acid containing 40%-propanol, and 0.4-ml fractions were taken with a fraction collector, FRAC-100 (Pharmacia Fine Chemicals). The fractions were checked for their CSA as described above and the active fractions were recovered. They were further purified on analytical u-Bondapak C18 column (4.6 mm x 30 cm), and the main peak was recovered and freeze-dried.
The protein so obtained was treated with 2-mercaptoethanol and subjected to Sps-polyacrylamide gel (15.0%) electrophoresis (15 mV, 6 hours). Upon staining with Coomassie Blue, the desired G-CSF polypeptide could be identified as a single band.
Example 17: Assay of G-CSF Activity (+VSE) The CSF sample obtained in Example 16 was assayed in accordance with the method of CSF assay (a) described earlier in this specification. The results are shown in Table 1.

-55- ~ ~ 4 ~ 3 B 9 Table 1 Human neutrophilic colonies (colonies/dish) Purified human G-CSF (20 ng) 73 CSF sample obtained in 68 Example 15 (50 ng) Blank 0 Example 18: Amino Acid Analysis (+VSE) 1) Analysis of amino acid composition The CSF sample purified in Example 16 was hydrolyzed by routine procedures, and the amino acid composition of the protein portion of the hydrolyzate was analyzed by a method of amino acid analysis with an automatic amino acid analyzer, Hitachi 835 (Hitachi Ltd.) The results are shown in Table 2.
Hydrolysis was conducted under the following conditions:
(l) 6 N HCI, 1.10°C, 24 hours, in vacuum (ii) 4 N methanesulfonic acid + 0.2% 3-(2-aminoethyl)indole, 110oC, 24 hours, 48 hours, 72 hours, in vacuum The sample was dissolved in a solution (1.5 ml) containing 40% n-propanol and 0.1% trifluoroacetic acid.
Aliquots each weighing 0.1 ml were dried with a dry nitrogen gas and, after addition of the reagents listed in (l) or (ii), the containers were sealed in vacuum, followed by hydrolysis of the contents.
Each of the values shown in Table 2 was the average of f our measurements, 24 hour value for (l) and 24, 48 and 72 hour values for (ii), except that the contents of Thr, Ser, 1/2 Cys, Met, Val, Ile arid Trp were calculated by the following methods (see "Tampaku Kagaku (Protein Chemistry) II", A Course in Biochemical Experiments, Tokyo Kagaku Dohjin):
-- For Thr, Ser, 1/2 Cys and Met, the time-dependent profile of the 24, 48 and 72 hour values for (ii) was extrapolated by zero hours.
-- For Val and Ile, the 72 hour value for (ii) was used.

-- For Trp, the average of 24, 48 and 72 hour values for (ii) was used.
Tab a 2 (Amino Acid Analysis Data) Amino acids Mole%

Asp (Asp + Asn) 2.3 Thr 4.0 Ser 8.5 Glu (Glu + Gln) 15.2 Pro 7.3 Gly 7.9 Ala ~ 10.7 ~

1/2 Cys 2.8 Val 4.5 Met 2.0 Tle 2.3 Leu 18.3 Tyr 1.7 Phe 3.4 Lys 2.3 His 2.8 Trp 1.1 Arg 2.9 2) Analysis of N-terminal amino acids The sample was subjected to Edman decomposition with a gas=phase sequences (Applied Biosystems) and the PTH amino acid obtained was analysed by routine procedures with a high-pressure liquid chromatographic apparatus (Beckman Instru-ments) and Ultrasphere-ODS column (Beckman Instruments).

1 3 ~+1 X89 After the column (5 ~..~m; 4.6 mm~ x 250 mm) was equilibrated with a starting buffer [an aqueous solution containing 15 mM
sodium acetate buffer (pH 4.5) and 40% acetonitrile], the sample (as dissolved in 20 u1 of the starting buffer) was injected and separation was achieved by isocratic elution with the starting buffer. During these operations, the flow rate was held at 1.4 ml/min and the column temperature at 40°C. Detection of the PTH amino acid was accomplished using the absorptions in the ultraviolet range at 269 nm and 320 nm. Standard samples (each weighing 2 nmol) of PTH
amino acid (Sigma) had been separated on the same line to determine their retention times, which were compared with those of the sample for the purpose of identification of the N-terminal amino acids. As a result, PTH-methionine and PTH-threonine were detected.
Example 19: Construction of E. coli Recombinant Vector (-VSE) and Transformation 1) Using tac promoter-containing vector The procedures of Example 12 were regeated except that the "pBRG4 prepared ~n Example 9 which contained the cDNA shown in Fig. 3(A)"~Isee (iii) in Example 12] was replaced by the "pBRV2 prepared in Example 10 which contained the cDNA shown in Fig. 4(A)". As in Example 12, the transformants obtained were verified as the desired ones (Fig. 11).
2) Using PL promoter-containing vector The procedures of Example 13 were repeated using cDNA
(-VSE) and the transformants obtained were verified as the desired ones (Fig. 12).
3) Using trp promoter-containing vector The procedures of Example 14 were repeated using cDNA
(-VSE) and the transformants were verified as the desired ones (Fig. 13).
Example 20: Assay of G-CSF Activity (-VSE) The three species of transformants obtained in Example 19 were cultured by the method described in Example 15. From the cultured E. coli cells, G-CSF polypeptides were recovered and purified by the method described in _$$_ Example 16, with the result that human G-CSF polypeptide was obtained as a single band.
The so obtained CSF sample was assayed by the method of CSF activity assay (a) described earlier in this specifi-cation. The results are shown in Table 3.
Table 3 Human neutrophilic colonies (colonies/dish) Purified human G-CSF (20 ng) 73 CSF sample obtained in 73 Example 19 (50 ng) Blank 0 Example 21: Amino Acid Analysis (-VSE) 1) Analysis of amino acid composition The amino acid composition of the CSF sample purified in Example 20 was analyzed by the method described in 1) in Example 18. The results are shown in Table 4.

1 ~ 41 389 2) Analysis of N-terminal amino acids The sample was subjected to analysis of the N-terminal amino acids in accordance with the method described in 2) in Example 18. As a result, PTH-methionine and PTH-threonine were detected.
Example 22: Preparation of pHGA410 Vector (for Use with Animal Cells, +VSE Line) The EcoRI fragment prepared in Example .9 which had the cDNA shown in Fig. ~(A) was treated with a restriction Tab a 4 (Amino Acid Analysis Data) 1 ~ 41 389 enzyme, DraI, at 37°C for 2 hours, followed by treatment with the Rlenow fragment of DNA polymerase I (Takara Shuzo Co., Ltd.) to create blunt ends. One microgram of BglII
linker (8mer, Takara Shuzo Co., Ltd.) was phosphorylated with ATP and joined to about 1 ug of the separately obtained mixture of DNA fragments. The joined fragments were treated with a restriction enzyme, BgIII, and subjected to agarose gel electrophoresis. Subsequently, only the largest DNA
fragment was recovered.
This DNA fragment was equivalent to about 710 base pairs containing a human G-CSF polypeptide coding portion (see Fig. 6). A vector pdKCR [Fukunaga et al., Proc. Natl.
Acad. Sci., USA, 81, 5086 (1904)) was treated with a restriction enzyme, BamHI, and subsequently dephosphorylated with an alkali phosphatase (Takara Shuzo Co., Ltd.) The vector DNA obtained was joined to the 710-by cDNA fragment in the presence of T4 DNA ligase (Takara Shuzo Co., Ltd.), so as to produce pHGA410 (Fig. 14). As shown in Fig. 14, this plasmid contained the promoter of SV4U early gene, the replication replication origin of SV40, part of the rabbit S-globin gene, the replication initiating region of pBR322 and the pBR322-derived g-lactamase gene (Ampr), with the human G-CSF gene being connected downstream of the promoter of the SV40 early gene.
Example 23: Construction of Recombinant Vector (+VSE) for Use in Transformation of C127 Cells 1) Construction of pHGA410 (H) Twenty micrograms of the plasmid pHGA410 (Fig. 14) prepared in Example 22 was dissolved in a reaction solution composed of 50 mM Tris-HC1 (pH 7.5), 7 mM MgCl2, 100 mM
NaCl, 7 mM 2-mercaptoethanol and 0.01 bovine serum alubmin (BSA). A restriction enzyme, EcoRI (10 - 15 units; Takara Shuzo Co., Ltd.) was added and the reaction solution was held at 37°C for about 30 minutes to cause partial digestion with EcoRI. Subsequently, the DNA fragment was subjected to two treatments with a 1:1 mixture of phenol/chlorof orm, one treatment with ether, and precipitation with ethanol.

The DNA fragment obtained was dissolved in 50 u1 of a solution composed of 50 mM Tris-HC1, 5 mM MgCl2, 10 mM DTT, and 1 mM each of dATP, dCTP, dGTP and dTTP. After 5 u1 of the Rlenow fragment of ,~ co i DNA polymerase (Takara Shuzo Co., Ltd.) was added, the solution was incubated at 14°C for 2 hours to produce blunt ends.
By subsequent 0.8% agarose gel electrophoresis, 6 up of a DNA fragment of about 5.8 kbp in length was recovered.
Five micrograms of the recovered DNA fragment was re dissolved in 50 u1 of a reaction solution composed of 50 mM
Tris-HC1 (pH 7.6), 10 mM MgCl2, 10 mM DTT and 1 mM ATP.
After 2 up of HindIIi linker (Takara Shuzo Co., Ltd.) and 100~units of T4 DNA lipase (Takara Shuzo Co., Ltd.) were added, reaction was carried out overnight at 4°C.
Subsequently, treatments with phenol and ether and precipitation with ethanol were conducted. The precipitate was dissolved in 30 ~l of a solution composed of 10 mM Tris-HC1 (pH 7.5), 7 mM MgCl2 and 60 mM NaCl, and the solution was incubated at 37°C for 3 hours in the presence of 10 units of HindIII. After re-treatment with T~ DNA lipase, the resulting DNA was used to transform Es c~li strain DH1 by the rubidium chloride procedure (see Molecular Cloning, ibid.) From ampicillin-resistant (Ampr) colonies of the transformants, cells were selected which harbored a plasmid which was identical to pHGA410 except that HindIII was inserted at the EcoRi site. The so obtained plasmid was named pHGA410 (H) (Fig. 15).
2) Construction of expression recombinant vector pTN-G4 Twenty micrograms of the pHGA~lO (H) thus obtained was dissolved in 50 u1 of a reaction solution composed of 10 mM Tris-HCl (pH 7.5), 7 mM MgCl2, 175 mM NaCl, 0.2 mM EDTA, 7 mM 2-mercaptoethanol and 0.01% bovine serum albumin.
After 20 units of Sall (Takara Shuzo Co., Ltd.) were added, the reaction solution was incubated at 37°C for 5 hours.
Following treatment with phenol and precipitation with ethanol, incubation was conducted as in 1) for about 2 hours at 14°C in the presence of the Rlenow fragment of DNA
polymerase (Takara Shuzo Co., Ltd.), so as to create blunt ends. Without being subjected to DNA recovery by agarose gel electrophoresis, the reaction solution was immediately subjected to precipitation with ethanol. The resulting DNA
fragment was treated with HindIII and 5 g of a HindIII-SalI
fragment (ca. 2.7 kbp) was recovered by 1~ agarose gel electrophoresis. In a separate step, a plasmid pdBPV-1 having a bovine papilloma virus (BPV) [this pl.asmid was obtained by courtesy of Dr. Howley and is described in Sarver, N, Sbyrne, J.C. & Howley, P.M., Proc. Natl. Acad.
Sci., USA, 79, 7147-7151 (1982)] was treated with HindIII
and PvuII, as described by Nagata et al. [Fukunaga, Sokawa and Nagata, Proc. Natl. Acad. Sci., USA, 81, 5086-5090 (1984)], to obtain an 8.4-kb DNA fragment. This 8.4-kb DNA
fragment and the separately obtained HindIII-SalI DNA
fragment (ca. 2.7 kb) were ligated by T4 DNA ligase. The ligation product was used to transform E. coli strain DH1 by the rubidium chloride procedure described in Molecular Cloning, ibid. E. coli colonies harboring a plasmid having the pHGA410-derived G-CSF'~cDNA were selected. This plasmid was named pTN-G4 (Fig. 15).
~sc~enovirus type II [Tanpakushitsu, Rakusan, Koso (Proteins, Nucleic Acids, and Enzymes), 27, December, 1982, Ryoritsu Shuppan] was similarly treated to obtain a plasmid ~ pVA, that contained a ca. 1700-by Sall-HindIII fragment harboring VAI and VAII, and.a fragment containing VAI and VAII was recovered from this plasmid. This fragment was inserted into pTNG4 at the HindIII site so as to obtain pTNG4VAa and pTNG4VA~ (Fig. 15). Because of the VA gene of adenovirus, these plasmids were capable of enhanced expres-sion of a transcription product from the early promoter of SV40.
Example 24: Transformation of C127 Cells and G-CSF
Expression Therein (+VSE) Before it was used to transform mouse 0127 cells, the pTN-G4 obtained in Example 23 was treated with a restriction enzyme, BamHI. Twenty micrograms of the plasmid pTN-G4 was dissolved in 100.u1 of a reaction solution [10 mM Tris-HC1 (pH 8.0), 7 mM MgCl2, 100 mM NaCI, 2 mM 2-mercaptoethanol ~ ~~~ 389 -63_ and 0.01$ BSA] and treated with 20 units of BamHI (Takara Shuzo Co., Ltd.), followed by treatments with phenol and ether, and precipitation with ethanol.
Mouse C127I cells were grown in a Dulbecco's minimal essential medium containing 10% bovine fetal serum (Gibco).
The C127I cells growing on plates (5 cm~) were transformed with 10 up, per plate, of the separately prepared DNA by the calcium phosphate procedure [see Haynes, J. & Weissmann, C., Nucleic Acids Res., 11, 687-706 (1983)]. After treatment with glycerol, the cells were incubated at 37oC for 12 hours.
The incubated cells were transferred onto three fresh plates (5 cm~) and the media were changed twice a week. At day 16, the foci were transferred onto fresh plates and subjected to serial cultivation on a Dulbecco's minimal essential medium containing 10% bovine fetal serum (Gibco), so as to select clones having high G-CSF production rate.
These clones produced G-CSF at a level of approximately 1 mg/L. Further cloning gave rise to clanes that were capable of producing G-CSF at levels of 10 mg/L or higher. In addition to the C127I cells, NIEi3T3 cells could also be used as host cells.
Example 25: Expression of G-CSF in CHO Cells (+VSE) 1) Construction of pHGG4-dhfr Twenty micirograms of the plasmid pHGA410 obtained in Example 22 was dissolved in 100 u1 of a reaction solution containing 10 mM Tris-HC1 (pH 7.5), 7 mM MgCl2, 175 mM NaCl, 0.2 mM EDTA, 0.7 mM 2-mercaptoethanol and 0.01% BSA. Reac-tion was carried out overnight at 37oC in the presence of 20 units of a restriction enzyme SalI (Takara Shuzo Co., Ltd.), followed by treatments with phenol and ether and precipita-tion with ethanol.
The precipitate of DNA was dissolved in 100 u1 of a reaction solution composed of 50 mM Tris-HC1, 5 mM MgCl2, 10 mM DTT, and 1 mM each of dATP, dCTP, dGTP and dTTP, and reaction was carried out at 14°C for 2 hours in the presence of the Klenow fragment of -E. coli DNA polymerase (10 u1;

Takara Shuzo Co., Ltd.), followed by treatments with phenol and ether, and precipitation with ethanol.
An EcoRI linker was attached to the DNA in the pre-cipitate by the following procedures: the DNA was dissolved in 50 u1 of a reaction solution composed of 50 mM Tris-HC1 (pH 7.4), 10 mM DTT, 0.5 mM spermidine, 2 mM ATP, 2 mM
hexamine-cobalt chloride and 20 ug/ml of BSA. Reaction was carried out at 4°C for 12 - 16 hours in the presence of EcoRI linker (Takara Shuzo Co., Ltd.) and 200 units of T4 DNA ligase (Takara Shuzo Co., Ltd.) After treatment with phenol, washing with ether and precipitation with ethanol, all being conducted in accordance with routine procedures, the DNA precipitate was partially digested with EcoRI and 3 ~g of a DNA fragment of about 2.7 kbp in length was recovered by 1~ agarose gel electroghoresis.
The plasmid pAdD26S'VpA [Kaufman, R.G. & Sharp, P.A., Mol. Cell Biol., _2, 1304-1319 (1982)] was treated with EcoRI
and dephosphorylated by treatment with a bacterial alkaline phosphatase (BAP). More specifically, 20 ug of pAdD26SVpA
and 20 units of EcoRI were added to a reaction solution [50 mM Tris-HC1 (pH 7.5), 7 mM MgGl2, 100 mM NaCl, 7 mM 2-mercaptoethanol and 0.01$ BSA] and reaction was carried out at 37°C for 10 hours. Subsequently, 5 units of BAP was added to the reaction solution, and reaction was carried out at 68oC for 30 minutes. Following treatment with phenol, the EcoRI fragment of pAaD26SVpA was recovered by electro-phoresis in a yield of approximately 5 fig.
The fragment of about 2.7 kbp in length and the pAaD26SVpA, each weighing 0.5 ug, were annealed. The resulting plasmid was used to transform E. coli strain DH1 by the rubidium chloride procedure, and the colonies harbor-ing the plasmid of pHGG4-dhfr were selected. The obtained plasmid was named pHGG4-dhfr (Fig. 16a).
The alternative procedure was as follows: the plas-mid pHGA410 was treated with SaII and partially digested with EcoRI without any EcoRI linker being attached. A DNA
fragment of about 2.7 kbp in length was recovered and treated with the Klenow fragment of k~. coli DNA polymerase ~ 3~1 389 to create blunt ends. An EcoRI fragment having blunt ends was prepared from pAaD26SVpA as described above. This EcoRI
fragment and the separately prepared fragment (ca. 2.7 kbp) were treated with T~ DNA ligase to prepare pHGG4-dhfr.
The pHGA410 (H) prepared in Example 23 was treated with restriction enzymes, HindIII and Sall, as described in 2) in Example 23, and the HindIII-SalI fragment was joined to the blunt-ended EcoRI fragment of pAdD26SVpA described above. This method could also be employed to prepare pHGG4-dhfr (Fig. 16b) .
2) Construction of pG4DR1 and pG4DR2 Ten micrograms of the plasmid pAaD26SVpA mentioned in 1) was dissolved in 50 ml of a reaction solution containing 50 mM Tris-HC1 (pH 7.5), 7 mM MgCl2, 100 mM NaCl, 7 mM 2-mercaptoethanol and 0.01% BSA. After addition of 10 units each of the restriction enzymes, EcoRT and BamHI, reaction was carried out at 37°C for 10 hours, followed by treatment with phenol and washing with ether. A DNA fragment of ca. 2 kb was recovered by electrophoresis through a 1% low-melting point agarose gel. The recovered DNA fragment was treated with the Klenow fragment of DNA polymerase by routine proce-dures so as to create blunt ends. The blunt-ended DNA frag-ment was subjected to treatment with phenol, washing with ether and precipitation with ethanol.
Ten micrograms of the plasmid pHGA410 (H) obtained in 1) of Example 23 was dissolved in SO u1 of a reaction solu-tion containing 10 mM Tris-HC1 (pH, 7.5), 7 mM MgCl2 and 60 mM NaCl. Reaction was carried out at 37°C for 6 hours in the presence of 10 units of HindIII. A DNA fragment was recovered by electrophoresis through a 1% low-melting point agarose gel that was conducted by routine procedures. The recovered DNA fragment was subsequently treated with BAP and blunt ends were created by treatment with the Rlenow frag-ment. Following treatment with phenol and washing with ether, the DNA fragment was joined at blunt ends to the previously obtained ca. 2-kb DNA fragment with a T4DNA

1 ~ 41 389 ligase by the following procedures: 1 ug of each DNA frag-ment was dissolved in 30 ul of a reaction solution contain-ing 66 mM Tris-HC1 (pH, 7.5), 6.6 mM MgCl2, 5 mM DTT and 1 mM ATP, and reaction was carried out at 6°C for 12 hours in the presence of 50 units of a T4DNA ligase. The ligation product was used to transform Es coli strain DHl. As a result, pG4DR1 and pG4DR2 shown in ~''ig. 16c were obtained.
3) Transformation and expression CHO cells (dhfr- strain; courtesy of Dr. L. Chasm of Columbia University) were cultivated for growth in alpha-minimal essential medium containing 10~ calf serum (a-MEN
supplemented with adenosine, deoxyadenosine and thymidine) in plates (9 cm~, Nunc). The cultured cells were trans-formed by the calcium phosphate procedure [Wigler et al., Cell, 14, 725 (1978)] in the following manner.
A carrier DNA (calf thymus DNA) was added in an appro-priate amount to 1 g of the plasmid pHGG4-dhfr prepared in 1), and the mixture was dissolved in 375 ul of a TE solution, followed by addition of 125 ul of 1 M CaCl2. After the solu-tion was cooled on ice for 3 - 5 minutes, 500 ul of 2 x HBS
(50 mM Hepes, 280 mM NaCl, and 1.5 mM phosphate buffer) was added to the solution. After re-coating on ice, the solu-tion was mixed with 1 ml of the culture of CHO cells, transferred onto plates, and incubated for 9 hours in a C02 incubator. The medium was removed from the plate and, following washing with TBS (iris-buffered saline), addition of 20$ glycerol-containing TBS, and re-washing, a non-selective medium (the a-MEN medium described above except that it was supplemented with nucleotides) was added. After 2-day incubation, a 10-fold dilution of the culture was transferred onto a selective medium (not supplemented with nucleotides). The cultivation was continued, with the medium being replaced by a fresh selective medium every 2 days, and the resulting colonies were selected and trans-ferred onto fresh plates, where the cells grew in the presence of 0.02 uM methotrexate (MTX), followed by cloning through growth in the presence of 0.05 uM MTX, which was later increased to 0.1 uM.

The transformation of CHO cells may also be accom-plished by cotransformation with pHGG4 and pAdD26SVpA (see Scahill et al., Proc. Natl. Acad. Sci., USA, 80, 4654-4658 (1983) ] .
CHO cells were also transformed by the following procedures: pG4DR1 or pG4DR2 that was prepared in 2) was preliminarily treated with Sall and KpnI respectively to obtain DNA fragments and 10 ug of these fragments was used to transform CHO cells as above; the transformed cells were subjected to continued cultivation in a series of selective media in the manner described above; about 7 days later, no less than 100 distinct colonies appeared per plates these colonies were transferred en masse to a fresh plate and subjected to continued cultivation in a series of selective media in the presence of 0.01 uM M'I'X, whereupon ten-odd colonies appeared; the same procedures were repeated with the MTX concentration being serially increased to 0.02 uM, 0.05 uM and 0.1 uM, and the colonies that survived were selected; colony selection.could be achieved in a similar manner even when the 10-odd colonies obtained were individ-ually selected and subj~bted to cultivation at increasing MTX concentrations.
A recombinant vector that harbors a "polycistronic gene" may also be used to transform CHO cells. An example of this alternative method is as follows: pAdD26SVpA was treated with PstI and the recovered two fragments were joined to a pBRG4-derived CSF cDNA fragment so as to construct a recombinant vector wherein the adeno virus promoter, CSF cDNA, DHFR and the poly(A) site of SV40 were inserted in the order written. This recombinant vector was used to transform CHO cells.
Example 26: Assay of G-CSF Activity (+VSE) The supernatants of cultures of 0127 cells and CHO
cells which were obtained in Examples 24 and 25, respec-tively, were adjusted to a pH of 4 with 1 N acetic acid.
After addition of an equal volume of n-propanol, the result-ing precipitate was removed by centrifugation. The super-natant was passed through an open column (l~ x 2 cmL) filled ~ 34~ 389 with a C8 reverse-phased carrier (Yamamura Kagaku R.K.) and elution was conducted with 50% n-propanol, The eluate was diluted two-fold with water and subjected to reverse-phased high-pressure liquid chromatography on YMC-C8 column (Yamamura Kagaku K.K.), followed by elution with n-propanol (30 - 60% linear density gradient) containing 0.1$ TFA. The fractions which were eluted at n-propanol concentrations of about 40% were recovered, freeze-dried and dissolved in 0.1 M glycine buffer (pH 9). As a result of these procedures, the human G-CSF in the C127 and CHO cells was concentrated about 20-fold.
As controls, cells were transformed with human G-CSF
cDNA-free plasmids and the supernatants of their cultures were concentrated in accordance with the procedures described above. The human G-CSF activities of the samples were assayed by the method of human G-CSF activity assay (a) described earlier in this specification. If the efficiency of expression is adequately high, the supernatants of cul-tures may be directly assayed without being concentrated.
The results are summarized in Table 5, wherein the data are based on concentrated samples.

Table 5 Assay of Human G-CSF Activity Human neutrophilic colonies (colonies/dish) Purified 96 human G-CSF
(20 ng) Culture of C127 cells transformed with pdBPV-1 0 (concentrated 20-f old}

Culture of 3T3 cells transformed with pdBPV-1 0 (concentrated 20-fold) B PV

Culture of 0127 cells transformed with pTNG4 82 (concentrated 20-fold) Culture of 3T3 cells transformed with pTNG4 85 (concentrated 20-fold) Culture of CHO cells transformed with pAdD26SVpA 0 (concentrated 20-fold}

Culture of CHO~~cells dhfr transformed with pHGG4-dhfr 110 (concentrated 20-fold) Culture of CHO cells transformed with pG4DRl 105 (concentrated 20-fold) Example 27: Amino Acid Analysis and Sugar Analysis (+VSE) 1) Analysis of amino acid composition The crude CSF sample~prepared in Example 26 was purified in accordance with the procedures described in Example 2(iii). The purified CSF sample was hydrolyzed by routine procedures, and the protein portion of the hydroly-zate was checked for its amino acid composition by a special method of amino acid analysis with Hitachi 835 automatic amino acid analyzer (Hitachi, Ltd.) The results are shown in Table 6. Hydrolysis was conducted under the following conditions:
(i) 6 N HC1, 110oC, 24 hours, in vacuum -70_ .(ii) 4 N methanesulfonic acid + 0.2% 3-(2-aminoethyl)indole, 110°~C, 24 hours, 48 hours, 72 hours, in vacuum.
The sample was dissolved in a solution (1.5 ml) containing 40% n-propanol and 0.1% trifluoroacetic acid.
Aliquots each weighing 0.1 ml were dried with a dry nitrogen gas and, after addition of the reagents listed in (i) or (ii), the containers were sealed in vacuum, followed by hydrolysis of the contents.
Each of the values shown in Table 6 was the average of f our measurements, 24 hour value for (i) and 24, 48 and 72 hour values for (ii), except that the contents of Thr, Ser, 1/2 Cys, Met, Val, Ile and Trp were calculated by the following methods (see "Tampaku Kagaku (Protein Chemistry) II", A Course in Biochemical Experiments, Tokyo Ragaku Dohjin):
-- For Thr, Ser, 1/2 Cys and Met, the time-dependent profile of the 24, 48 and 72 hour values for (ii) were extrapolated for zero hours.
-- For Val and Ile, the 72 hour value for (ii) was used.
-- For Trp, the average of 24, 48 and 72 hour values for (ii) was used.

_71_ .1 ~ 4 1 3 8 9 Table Amino Acid Analysis Data Amino acids ~ Mole$
Asp (Asp -w Asn ) 2. 3 Thr 3,9 Ser 8.5 Glu (Glu + G1n) ~ 15.3 Pro ~ 7.4 g t ly 7.8 Ala ' 10.8 1/2 Cys 2.8 Val 4.5 Met 1.7 Ile 2.3 Leu 18.6 Tyr 1.7 Phe 3.4 Lys 2. 3 His 2.8 Trp 1.1 Arg ~ 2.8 2) Sugar composition analysis An internal standard (25 nmal of inositol) was added to 200 ng of the purified CSF sample used in the analysis of amino acid composition l). After addition of a methanol solution (500 u1) containing 1.5 N HCI, reaction was carried out at 90°C for 4 hours in a N2 purged, closed tube. After the tube was opened, silver carbonate (Ag2C43) was added to neutralize the contents. Thereafter, 50 u1 of acetic ~ 3 41 389 anhydride was added and the tube was shaken for an adequate period. Subsequently, the tube was left overnight in the dark at room temperature. The upper layer was put into a sample tube and dried with a nitrogen gas. Methanol was added to the precipitate and the mixture was washed and lightly centrifuged. The upper layer was put into the same sample tube and dried. After addition of 50 1~1 of a TMS
reagent (5:1:1 mixture of pyridine, hexamethyl disilazane and trimethylchlorosilane), reaction was carried out at 40°C
for 20 minutes and the reaction product was stored in a deep freezer. A standard was prepared by combining 25 nmol of inositol with 50 nmol each of galactose (Gal), N-acetyl galactosamine (Gal NAc), sialic acid and any other appro-priate reagents.
The samples thus prepared were subjected to gas chromatographic analysis under the following conditions:
Conditions of analysis Column . 2$ OV - 17 VINgort HP, 60 - 80 mesh, 3 m, glass Temperature : elevated from 110 to 250°C at 4°C/min.
Carrier gas (N2) pressure . initially 1.2 - 1.6 kg/cm2 finally 2 - 2.5 kg/cm2 Sensitivity : 103 MS2 range, 0.1 - 0.4 volts Pressure : H2, 0.8 kg/cm2 air, 0.8 kg/cm2 Sample feed : 2.5 - 3.0 u1.
As a result of the analysis, galactose, N-acetyl galactosamine and sialic acid were identified in the CSF
sample of the present invention.
Example 28: Preparation of pHGV2 Vector (for Use with Animal Cells, -VSE line) The EcoRI fragment prepared in Example 10 which had the cDNA shown in Fig. 4(A) was treated with a restriction enzyme, DraI, at 37°C for 2 hours, followed by treatment with the Klenow fragment of DNA polymerase T (Takara Shuzo Co., Ltd.) to create blunt ends. One microgram of BglII
linker (8mer, Takara Shuzo Co., Ltd.) was phosphorylated with ATP and joined to about 1 ug of the separately obtained mixture of DNA fragments. The joined fragments were treated with a restriction enzyme, BgIII, and subjected to agarose gel electrophoresis. Subsequently, only the largest DNA
fragment was recovered.
This DNA fragment was equivalent to about 700 base pairs containing a human G-CSF polypeptide coding portion (see Fig. 6). A vector pdKCR [Fukunaga et al., Proc. Natl.
Acad. Sci., USA, 81, 5086 (1984)] was treated with a restric-tion enzyme, BamHI, and subsequently dephosphorylated with an alkali phosphatase (Takara Shuzo Co., Ltd.), The vector DNA obtained was joined to the about 700 cDNA fragment in the presence of T4 DNA ligase (Takara Shuzo Co., Ltd.), so as to produce pHGV2 (Fig. 17). As shown in Fig. 17, this plasmid contained the promoter of SV40 early gene, the replication initiating region of SV40, part of the rabbit S-globin gene, the replication initiating region of pBR322 and the pBR322-derived S-lactamase gene (Ampr), with the human G-CSF gene being connected downstream of the promoter of the SV40 early gene.
Example 29: Construction of Recombinant Vector (-VSE) for Use in Transformation of 0127 Cells 1) Construction of pHGV2(H) Twenty micrograms of the plasmid pHGV2 (Fig. 17) prepared in Example 28 was treated by the procedures described in 1) in Example 23, so as to prepare a plasmid named pHGV2(H) (Fig. 18).
2) Construction of expression recombinant vectors pTN-V2, pTNVAa and pTNVA~
With 20 ug of the pHGV2(H) being used, the procedures described in 2) in Example 23 were repeated to select E.
coli harboring a plasmid having the pHGV2-derived G-CSF
cDNA. This plasmid was named pTN-V2 (Fig. 18).
Adenovirus type II [Tampakushitsu, Kakusan, Koso (Proteins, Nucleic Acids, and Enzymes), 2,-7, December, 1982, Ryoritsu Shuppan] was similarly treated to obtain a plasmid, OpVA, that contained a ca. 1700-by SalI-HindIII fragment harboring VAI and VAII, and a fragment containing VAI and VAII was recovered from this plasmid. This fragment was ~ ~ 4~ 3a9 inserted into pTN-V2 at the HindIII site so as to obtain pTNVAa and pTNVAs (Fig. 18). Because of the VA gene of adenovirus, these plasmids were capable of enhanced expres-sion of a transcription product from the early promoter of SV40.
Example 30: Transformation of C127 Cells and G-CSF
Expression Therein (-VSE) The pTN-V2 obtained in Example 29 was treated with a restriction enzyme, BamHI, before it was used to transform mouse C127 cells.
Mouse C127I cells were transformed with the so prepared DNA to express G-CSF (see Example 24) and clones having high G-CSF production rate were selected. These clones produced G-CSF at a level of approximately 1 mg/L.
By further cloning, clones capable of producing G-CSF
at a level of 10 mg/L could be selected. In a similar manner, C127 cells were transformed with the pTNVAa and pTNVAB obtained in Example 29, and the transformants were selected for clones having high capability of G-CSF produc-tion; as for pTNVAa, clones capable of producing G-CSF at yields of 20 mg/L or more could be obtained, while clones having a lower productivity (a few mg/L) were obtained by transformation with pTNVA~.
In addition to the C127I cells, NIH3T3 cells could also be used as host cells.
Example 31: Expression of G-CSF in CHO Cells (-VSE) 1) Construction of pHGV2-dhfr A DNA fragment of about 2.7 kbp in length was prepared from 20 ug of the plasmid pHGV2 (Example 28) by the procedures described in 1) in Example 25. This fragment (0.5 ug) and the EcoRI fragment of pAaD26SVpA (0.5 ug) were annealed. The resulting plasmid was used to transform Es coli strain DH1 by the rubidium chloride procedure, and the colonies harboring the plasmid of pHGV2-dhfr were selected.
The obtained plasmid was named pHGV2-dhfr (Fig. 19a).
The alternative procedure was as follows: the plas-mid pHGV2 was treated with SalI and partially digested with ' 1 3 41 '~89 EcoRI without any EcoRI linker being attached. A DNA frag-ment of about 2.7 kbp in length was recovered and treated with the Rlenow fragment of E. coli DNA polymerase to create blunt ends. A blunt-ended EcoRI fragment was prepared from pAdD26SVpA as described above. This EcoRI fragment and the separately prepared fragment (ca. 2.7 kbp) were treated with T4 DNA ligase to prepare pHGV2-dhfr.
The pHGV2 (H) prepared in 1) of Example 29 was treated with restriction enzymes, HindIII and SalI, as described in 2) in Example 29, and the HindIII-SalI fragment was joined to the blunt-ended EcoRI fragment of pAdD26SVpA
described above. This method could also be employed to prepare pHGG4-dhfr (Fig. 19b).
2) Construction of pV2DRl and pV2DR2 Ten micrograms of the plasmid pAdD26SVpA mentioned in 1) was dissolved in 50 ml of a reaction solution containing 50 mM Tris-HCl (pH, 7.5), 7 mM MgCl2, 100 mM NaCl, 7 mM 2-mercaptoethanol and 0.01% BSA. Reaction was carried out at 37°C for 10 hours in the presence of 10 units each of the restriction enzymes, EcoRI and BamHI. Therefore, treatment with phenol and washing with ether were conducted by routine procedures. A DNA fragment of ca. 2 kb was recovered by electrophoresis through a 1% low-melting point agarose gel.
The recovered DNA fragment was treated with the Rlenow fragment of DNA polymerase by routine procedures so as to create blunt ends. The blunt-ended DNA fragment was subjected to treatment with phenol, washing with ether and precipitation with ethanol.
Ten micrograms of the plasmid pHGV2(H) obtained in 1) of Example 29 was dissolved in 50 ~l of a reaction solution containing 10 mM Tris-HC1 (pH, 7.5), 7 mM MgCl2 and 60 mM
NaCl. Reaction was carried out at 37°C for 6 hours in the presence of 10 units of HindIII. A DNA fragment was recovered by electrophoresis through a 1% low-melting point agarose gel that was conducted by routine procedures. The recovered DNA fragment was subsequently treated with BAP and blunt ends were created by treatment with the Klenow frag-ment. Following treatment with phenol and washing with ~ ~ 4~ 3a'~

ether, the DNA fragment was joined at blunt ends to the previously obtained ca. 2-kb DNA fragment with a T4DNA
lipase by the following procedures: 1 up of each DNA frag-ment was dissolved in 30 u1 of a reaction solution contain-s ing 66 mM Tris-HC1 (pH, 7.5), 6.6 mM MgCl2, 5 mM DTT and 1 mM ATP, and reaction was carried out at 6°C for 12 hours in the presence of 50 units of a T4DNA lipase. The ligation product was used to transform E. coli strain DH1. As a result, pV2DRl and pV2DR2 shown in fig. 19c were obtained.
3) Transformation and expression CHO cells were transformed with the plasmid pHGV2-dhfr for G-CSF expression in accordance with the procedures described in 3) in Example 25.
The transformation of CHO cells may also be accom-plished by cotransformation with pHGV2 and pAdD26SVpA.
CHO cells were also transformed by the following procedures: pV2DRl or pV2DR2 that was prepared in 2) was preliminarily treated with SalT and KpnI respectively to obtain DNA fragments and 10 up of these fragments was used to transform CHO cells as above; the transformed cells were subjected to continued cultivation in a series of selective media in the manner described above; about 7 days later, no less than 100 distinct colonies appeared per plate; these colonies were transferred en masse to a fresh plate and subjected to continued cultivation in a series of selective media in the presence of 0.01 uM MTX, whereupon ten-odd colonies appeared; the same procedures were repeated with the MTX concentration being serially increased to 0.02 uM, 0.05 uM and 0.1 uM, and the colonies that survived were selected; colony selection could be achieved in a similar manner even when the 10-odd colonies obtained were individ-ually selected and subjected to cultivation at increasing MTX concentrations.
A recombinant vector that harbors a "polycistronic gene" may also be used to transform CHO cells. An example of this alternative method is as follows: pAdD26SVpA was treated with Pstl and the recovered two fragments were joined to a pBRV2-derived CSF cDNA fragment so as to _77_ construct a recombinant vector wherein the adeno virus promoter, CSF cDNA, DHFR and the poly(A) site of SV40 were inserted in the order written. This recombinant vector was used to transform CHO cells.
Example 32: Assay of G-CSF Activity (-VSE) By the procedures described in Example 26, human G-CSF was obtained from the supernatants of cultures of C127 cells and CHO cells which were obtained in Examples 30 and 31, respectively. The human G-CSF activity of each of the recovered samples was assayed as in Example 26. The results are shown in Table 7, 1 34~ 389 Tab a 7 Assay of Human G-CSF Activity Human neutrophilic colonies (colonies/dish) Purified 96 human G-CSF
(20 ng) Culture of C127 cells transformed with pdBPV-1 0 (concentrated 20-fold) Culture of 3T3 cells transformed with pdBPV-1 0 (concentrated 20-fold) B PV

Culture of C127 cells transformed with pTN-V2 107 (concentrated 20-fold) Culture of 3T3 cells transformed with pTN-V2 103 (concentrated 20-f old) Culture of CHO cells transformed with pAdD26SVpA 0 (concentrated 20-fold) Culture of CHO~'cells dhfr transformed with pHGV2-dhfr 111 (concentrated 20-fold) s Culture of CHO cells transformed with pV2DR1 F

(concentrated 20-fold) Example 33: Amino Acid Analysis and Sugar Analysis (-VSE) 1) Analysis of amino acid composition The crude CSF sample 'prepared in Example 32 was purified in accordance with the procedures described in Example 2(iii). The purified CSF sample was subjected to analysis of amino acid composition by the procedures described in 1) in Example 27. The results are shown in Table 8.

Table 8 Amino Acid Analysis Data Amino acids Mole%

Asp (Asp + Asn) 2.3 Thr 4.0 Ser 8.1 Glu (Glu + Gln) 15.1 Pro 7.5 Gly 8.0 Ala 10.9 1/2 Cys 2.8 Val 3.9 Met 1.7 Ile 2.3 Leu 18.9 Tyr 1.7 Phe 3.5 Lys 2.3 s His ~ 2.9 c Trp ' 1.2 Arg ~ 2.9 2) Analysis of sugar composition The purified CSF sample used in the analysis of amino acid composition in 1) was also subaected to analysis of its sugar composition by the same procedures and under the same conditions as those described in 2) in Example 27. As a result of this analysis, the presence of galactose, N-acetyl galactosamine and sialic acid in the CSF sample of the present invention was verified.

1 34~ 389 _80-Example 34: Construction of Recombinant Veetar Containing Chromosomal Gene for Expression in COS Cells The plasmid pBRCE38 that was ok~tained in Example 11 and which contained the chromosomal gene shown in Fig. 5 was treated with EcoRI. The pSVH*K+ plasmid described by Banerji et al. in Cell, 27, 299 (1981) was treated with KpnI
to remove the globin gene. The plasmid was further subjected to partial digestion with HindIII so as to remove part of the late gene of SV40. The fragments were rejoined l0 to prepare an expression vector pML-E+.
This vector was treated with the restriction enzyme, EcoRI, and dephosphorylated with an alkaline phosphatase (Takara Shuzo Co., Ltd.) to obtain a vector DNA, which was linked to the aforementioned chromosomal DNA with the aid of a T4DNA ligase (Takara Shuzo Co., Ltd.) to obtain pMLCE3o(.
As shown in Fig. 20, this plasmid contained the enhancer of SV40 gene, the replication origin of SV40, the replication origin of pBR322 and the pBR322-derived ~-lactamase gene (Ampr), and had the human G-CSF chromosomal gene joined downstream from the enhancer of SV40 gene.
Example 35: Expression of Human G-CSF Chromosomal Gene in COS Cells COS-1 cells (provided by courtesy of Dr. Gluzman of Cold Spring Harbor Laboratory, D.S.A.) that had been grown to a density of about 70~ in Petri dishes (9 cm~, Nunc) using a DMEM medium (Dulbecco's mod~.fied Eagle's medium available from Nissui Seiyaku K.K. under the trade mark NISSUI > containing 10~ calf serum were transformed by either the calcium phosphate procedure [Wigler et al., Cell, 14, 725 (1978)) or the DEAF-dextran:chloroquine method [see, for example, Gordon et al., Science, 228, 810 (10985)].
Transformation by the calcium phosphate procedure was conducted as follows: 160 ug of the plasmid pMLCE3~x prepared in Example 34 was dissolved in 320 u1 of a TE
solution and, after addition of distilled water (3.2 ml), 504 u1 of 2 M CaCl2 was added.
To the resulting solution, 4 ml of 2 x HBS (50 mM
Hepes, 280 mM NaCl, 1.5 mM phosphate buffer, pH 7.12) was added and the mixture was cooled on ice for 20 - 30 minutes.
The cooled mixture was added dropwise to the medium in an amount of 1 ml per Petri dish where the COS-1 cells had grown. After cultivation for 4 hours at 37°C in a C02 incu-bato n the cells were washed with a serum-free DMEM medium, then left to stand for about 3 minutes at room temperature in 5 ml of a DMEM medium containing 20% glycerol, and re-washed with a serum-free DMEM medium. After the serum-free DMEM medium was removed, 10 ml of a DMEM medium containing 10% calf serum was added and cultivation was conducted over-night in a C02 incubator. After the medium was replaced by a fresh one of the same type, cultivation was conducted for an additional 3 days.
Transformation by the DEAE-dextran:chloroquine method was conducted as follows: as in the calcium phosphate procedure, COS-1 cells were cultivated to grow to a density of 70% and washed twice with a serum-free DMEM medium; to the washed cells, a serum-free DMEM medium containing 250 ug/ml of DEAF-dextran and 2 ug/ml of the plasmid pMLCE3o( prepared in Example 34 was added and cultivation was conducted at 37°C for 12 hours; subsequently, the cells were washed twice with a serum-free DMEM medium and subjected to further cultivation at 3?oC far 2 hours in a DMEM medium containing 10% calf serum and 1 mM chloroquine; thereafter, the cells were washed twice with a serum-free DMEM medium and cultured at 37°C for an additional 3 days in a DMEM
medium containing 10% calf serum.
The supernatant of the so obtained culture of COS-1 cells was adjusted to a pH of 4 with 1 N acetic acid. After addition of an equal volume of n-propanol, the resulting precipitate was removed by centrifugation. The supernatant was passed through an open column (l~ x 2 cmL) filled with a C8 reverse-phased carrier (Yarnamura Ragaku R.K.) and elution was conducted with 50% n-propanol. The eluate was diluted two-fold with water and subjected to reverse-phased high-pressure liquid chromatography on YMC-C8 column (Yamamura Ragaku R.R.), followed by elution with n-propanol (30 - 60%
linear density gradient) containing 0.1% TFA. The fractions 1 3 41 3~9 which were eluted at n-propanol concentrations of about 40%
were recovered, freeze-dried and dissolved in 0.1 M glyci-dine buffer (pH 9). As a result of these procedures, the human G-CSF in the supernatant of the culture of COS-1 cells was concentrated about 20-fold.
As controls, COS-1 cells were transformed with G-CSF
chromosomal-gene free pML-E+ by the above-described proce-dures and the supernatant of the resulting culture was concentrated.
The human G-CSF activities of the obtained samples were assayed by the "Method of Human G-CSF Activity Assay (a)" described earlier in this specification. The results are Summarized in Table 9.
Table 9 Human neutrophilic colonies (colonies/dish) Purified human G-CSF (20 ng) 18 Culture of COS cells ~

transformed with pML-E 0 (concentrated 20-fold) Culture of COS cells transformed with pMLCE3a 23 (concentrated 20-fold) Culture of COS cells transformed with pMLCE3a 19 (concentrated ZO-f old) Example 36: RNA Analysis of G-CSF (Chromosomal Gene) COS cells cultivated to a cell concentration of 8 x 106 cells/plate (9 cm~) were transformed with 80 ug of the plasmid pMLCE3a. After 48 hours, the totel RNA was prepared in accordance with the procedure of Chirgwin [Biochemistry, 18, 5294 - 5299 (1979)x.
The plasmid pBRG4 obtained in Example 9 was cleaved with restriction enzyme AhaIII and the resulting pBRG4-derived DNA fragment was radiolabelled with [Y-32P]ATP
using T4 polynucleotide kinase to obtain an ca. 2.8-kb DNA
fragment containing G-CSF cDNA. The fragment was recovered and used as a DNA probe. After the DNA probe (I.5 x 105 c.p.m., 2.8 x 106 c.p.m./ug DNA) was denatured, it was mixed with 20 ~,g of the total RNA prepared from COS cells.
Hybridization at 45°C for 15 hours was conducted. The mixture was digested with 200 units/ml or A00 units/ml of S1 nuclease (P.L. Biochemicals) in accordance with the procedures of Weaver and Weissmann [Nucleic Acid Res., 7,~
1175 - 1193 (1979)], followed by 4% polyacrylamide gel electrophoresis in the presence of 8.3 M urea. Detection by autoradiography was then conducted.
As a result, a band corresponding to 722 by was observed as a strongly radiolabelled band in COS cells, from which a band corresponding to 487 by was also detected.
Therefore, the RNA of COS cells was found to contain G-CSF mRNAs of both +VSE and -VSE line.
Example 37: Amino Acid Analysis and Sugar Analysis (Chromosomal Gene) 1) Analysis of amino acid composition The crude CSF sample.prepared in Example 35 was purified in accordance with the procedures described in Example 2(iii). The purified CSF sample was subjected to analysis of amino acid composition by the procedures described in 1) in Example 27. The results are shown in Table 10.

~ 3 41 389 -s4-Table 10 Amino Acid Analysis Data Amino acids ~ Mole%
Asp (Asp + Asn) 2.3 Thr 4.9 Ser 8.3 Glu (Glu + Gln) 15.3 Pro 7.4 Gly 7.9 Ala 10.8 1/2 Cys 2.8 Val , 4.3 Met 1.7 Ile ~ 2.
i 3 Leu 18.7 Tyr ~ 1.7 Phe = 3.4 Lys ' 2.3 His ~ 2.9 Trp ' 1.1 f Arg ; 2.9 2) Analysis of sugar composition The purified CSF sample used in the analysis of amino acid composition in 1) was also subjected to analysis of its sugar composition by the same procedures and under the same conditions as those described in 2) in Example 27. As a result of this analysis, the presence of galactose, N-acetyl galactosamine and sialic acid in the CSF sample of the present invention was verified.

Example 38: Expression of Human G-CSF Chromosomal Gene in C127 Cells The plasmid pMLCE3a obtained in Example 34 was treated with EcoRI and a fragment of ca. 4 kb was recovered by the procedures described in Molecular Cloning, ibid. The recovered fragment was used as a source of the chromosomal G-CSF gene.
The fragment was treated with the Rlenow fragment of DNA polymerise I to create blunt ends (A).
The promoter of SV40 (ca. 0.4-kb EcoRI-EcoRI
fragment) was cut out from the plasmid pHGA410 (as prepared in Example 22) by the procedures described in Molecular Cloning, ibid., and was subsequently treated with the Klenow fragment of DNA polymerise (B).
' In a separate step, a plasmid pdBPV-1 having a bovine papilloma virus (BPV) [this plasmid was obtained by courtesy of Dr. Howley and is described in Sarver, N., Sbyrne, J.C. &
Howley, P.M., Proc. Natl. Acid. Sci., USA, 79, 7147-7151 (1982)] was treated with HindTII and PvuII to obtain a DNA
fragment of ca. 8.4 kb. This fragment was treated with the Klenow fragment of DNA polymerise T and dephosphorylated with a bacterial alkaline phosphatase (C).
The DNA fragments (A), (B) and (C) each weighing 0.1 ug were dissolved in 20 u1 of a reaction solution [50 mM
Tris-HC1 (pH 7.6), 10 mM MgCl2, 10 mM DTT, 1 mM ATP] and reaction was carried out overnight at 4°C in the presence of 180 units of a T4DNA ligase.
The reaction solution was subsequently treated by the rubidium chloride procedure described in Molecular Cloning, ibid. so as to obtain the plasmid pTNCE3a (Fig. 21).
The DNA fragment (A) used as a source of the chromo-somal G-CSF gene may be replaced by a DNA fragment of ca.
1.78 kb that is obtained by the following procedures: 20 ug of pMLCE3a is dissolved in 100 u1 of a mixture of 10 mM
Tris-HC1 (pH 8.0), 7 mM MgCl2, 100 mM NaCl, 7 mM 2-mercaptoethanol and 0.01% BSA; the solution is incubated at 37oC for 5 hours in the presence of 20 units of StuI and subjected to electrophoresis through 1.2% agarose gel.

f The so obtained plasmid pTNCE3a was used to transform mouse C127 I cells as in Example 24 and clones that ex-pressed the human G-CSF chromosomal gene and which had a high capacity for producing G-CSF were selected.
Example 39: Expression of Human G-CSF Chromosomal Gene in CHO Cells As in the case of expression in C127 cells, the plasmid pMLCE3a was treated with Stul and a DNA fragment of ca. 1.78 kb was recovered; alternatively, the same plasmid was treated with EcoRI and an EcoRI fragment of about 4 kb was recovered. Either fragment was suitable for use as a source of the chromosomal G-CSF gene.
The source fragment was treated with the Rlenow fragment of DNA polymerase I (a).
As in Example 38, the gromoter of SV40 (EcoRI-EcoRI
fragment) was cut out from pHGA410 to obtain a fragment of about 0.4 kb, which was similarly treated with the Rlenow fragment of DNA polymerase (b).
In a separate step, the plasmid pAdD26SVpA plasmid [Kaufman, R.G. & Sharp. P.A.. Mol. Cell. Biol., 2, 1304-1319 (1982)] was treated with EcoRI, then with the Klenow frag-ment of DNA polymerase, and finally dephosphorylated by treatment with a bacterial alkaline phosphatase (c).
The fragments, (a), (b) and (c), each weighing 0.1 ug were dissolved in 20 ~1 of a reaction solution [50 mM Tris-CH1 (pH 7.6), 10 mM MgCl2, 10 mM DTT, 1 mM ATP] and reaction was carried out overnight at 4°C in the presence of 180 units of a T4DNA ligase.
The reaction solution was subsequently treated by the rubidium chloride procedure described in Molecular Cloning, ibid., so as to transform 1E. coli strain DH1. The resulting Tetr colonies were screened for those containing the plasmid pD26SVCE3a.
As shown in Fig. 22, the plasmid pD26SVCE3a has the CSF gene linked to the early gene of SV40, and the dhfr gene linked downstream from the principal late promoter of adenovirus.

~ 3 ~~ 3ss The plasmid pAdD26SVpA was treated with EcoRI and BamHI as in 2) of Example 25, so as to obtain a DNA fragment (ca. 2 kb) containing the dhfr gene. This fragment was linked to fragment (a) and the EcoRI-SaII fragment of pHGA410 (H), so as to construct an Ampr expression vector pDRCE3a (Fig. 22).
CHO cells were transformed with the so obained plasmids, pD26SVCE3a and pDRCE3a, as in Example 25. By repeated selection through growth in the presence of MTX, clones of a G-CSF producing strain were obtained.
Example 40: Assay of the G-CSF Activity of Transformants (expressing human chromosomal gene) The supernatants of cultures of 0127 cells and CHO
cells which were obtained in Examples 38 and 39, respec-Lively, were worked up as in Example 26 to obtain human G-CSF and its activity was assayed. The results are shown in Table 11.

-~~- 1 ~ 41 389 Table 11 Assay of Human G-CSF Activity Human neutrophilic colonies (colonies/dish) Purified 85 human G-CSF
(20 ng) Culture of C127 cells transformed with pdBPV-1 0 (concentrated 20-fold) B PV

Culture of 0127 cells transformed with pTNCE3a 83 (concentrated 20-fold) Culture of CHO cells transformed with pAdD26SVpA 0 (concentrated 20-fall) Culture of CHO cells dhfr transformed with pD26SVCE3a 85 (concentrated 20-fold) Culture of CHO cells transformed with- pDRCE3a 86 (concentrated 20-fold) Example 41: Molecular Weight and Isoelectric Point of Transf ormants The gurified CSF samples used in the analysis of amino acid composition in Examples 16, 20, 27, 33 and 37 were subjected to measurements of their molecular weights and isoelectric points by the following procedures.
1) Molecular weight The molecular weight of the CSF was determined by sodium dodecylsulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The electrophoretic equipment was PROTEANTH
(16 cm, product of Bio-Rad Corporata.on), using a gel made up of a polyacrylamide slab gel (T = 15%, C = 2.6%) measur-ing 140 mm x 160 mm x 1.5 mm, and a concentrating gel (T =
3%, C = 20%). A denatured CSF sample was prepared by the following procedure: CSF was boiled for 3 minutes in a solution containing 2% of sodium dodecylsulfate in 0.46 M 2-mercaptoethanol. After performing electrophoresis with 4 u9 of the sample with a constant current of 30 mA for 4 hours, ~ 3 41 389 -89_ the gel was removed and stained with 0.25% Coomassie Brilliant Blue R 250 (product of Sigma Chemical Co.) for band detection. The following substances were used as molec-ular weight markers after similar treatments: phosphorylase B (mol. wt. 92,500), bovine serum albumin (BSA, mol. wt.
67,000), ovalbumin (OVA, mol, wt. 45,000), carbonic anhydrase (mol, wt. 31,000), soybean trypsin inhibitor (mol. wt. 21,500) and lysozyme (mol wt. 14,400).
As a result, a single band corresponding to a molec-ular weight of 185,000 ~ 1,000 was detected for each of the CSF samples obtained in Example 16 [E. coli/cDNA (+VSE)]
and Example 20 [E. coli/cDNA (-VSE)], and a single band correponding to a molecular weight of 19,000 + 1,000 was detected from each of the CSF samples obtained in Example 27 [C127,CH0/cDNA (+VSE)], Example 33 ~:C127,CH0/cDNA (-VSE)]
and Example 37 (COS/gDNA).
2) Isoelectric point The isoelectric point of the CSF of the present invention was determined by a flat bed, isoelectric electro-phoretic apparatus, FBE-3000 (product of Pharmacia Fine Chemicals). After 2-hour electrophoresis with a constant power of 30 watts (Vmax = 2,000 volts) on a polyacrylamide gel (T = 5%, C = 3%, 115 mm x 230 mm) containing Pharmalyte (pH = 4 - 6.5, Pharmacia Fine Chemicals) and 4M urea, the CSF was fixed with 30% methanol/10% trichloroacetic acid/35%
sulfosalicylic acid, and stained with Coomassie Brilliant Blue R-250. A Low pI kit (pH: 2.5 - 6.5, product of Pharmacia Fine Chemicals) was used as a isoelectric point marker.
Analysis of band separation at a pH of 4 to 6.5 gave a single band corresponding to pI = 6.1 for each of the CSF
samples obtained in Example 16 and 2U, and gave three distinct bands corresponding to pI = 5.5, 5.8 and 6.1 for each of the CSF samples s~ obtained in Example 27, 33 and 37.
Example 42: Protective Effect of Human G-CSF against Microbial Infection Test Method 1. Protection against infection with Pseudpmonas aeruginosa ~ 34~ ~a9 -90_ Endoxan (trade name of Shionogi & Co., Ltd.) was administered intraperioneally into 8-9-wk-old ICR mice (male:
35.3 ~ 1.38 g in body weight) in a dose of 200 mg/kg. The mice were then divided into three groups; two groups were given four subcutaneous injections (each 0.1-ml dose), at 24-hr intervals, of a solvent [1$ propanol and 0.5$ (w/v) mouse serum albumin in physiological saline] containing human G-CSF
(25,000 or 50,000 units/mouse), whereas the other group was given only the solvent in accordance with the same schedule.
Three hours after the last injection, the mice in each group were infected with Pseudomonas aeruginosa GNB-139 by subcutaneous injection (3.9 x 105 CFU/mouse). Twenty-one hours after the infection, the first two groups were given another subcutaneous injection of the solvent containing human G-CSF (25,000 or 50,000 units/mouse) and the other group given the solvent only.
The protective effect of human G-CSF was checked by counting the number of mice which were alive 10 days after the infection.
Preparation of cell suspension Pseudomonas aeru inosa GNB-139 was cultured overnight with shaking at 37°C in a Heart Infusion liquid medium (trade name of Difco}. The culture was suspended in a physiological saline solution.
2. Protection against infection with Candida Endoxan (trade name of Shionogi & Co., Ltd.) was administered intraperitoneally into 8-wk-old ICR mice (male;
40.5 ~ 1.60 g in body weight) in a dose of 200 mg/kg. The mice were then divided into two groups; one group was given four subcutaneous injections (each 0.1-ml dose), at 24-hr intervals, of a solvent [1~ propanol and 10$ (w/v) ICR mouse serum in physiological saline] containing human G-CSF
(50,000 units/mouse), whereas the other group was given only the solvent in accordance with the same schedule. Four hours after the last injection, the mice in each group were infected with Candida albicans U-50-1 (strain isolated from urine of leukemic patients; courtesy by Bacteriological 1 34~ 3$9 -91_ Laboratory, Tohoku University, School of Medicine) by intra-venous injection (5.6 x 105 CFU/mouse). The protective effect of human G-CSF was checked by counting the number of mice which were alive ten days after the infection.
Preparation of cell suspension Candida albicans U-50-1 was cultured overnight with shaking at 37°C in a yeast extract-containing Sabouraud liguid medium (2~ dextrose from Junsei Pure Chemicals Co., Ltd.; 10$ Tryptocase Peptone, trade name of BBL; 5$ yeast extract from Difco; pH, 5.6). The culture was washed twice with physiological saline and suspended in physiological saline.
3. Protection against infection with intracellular parasitic Listeria Endotoxan (trade name of Shionogi & Co., Ltd.) was administered intraperitoneally to 7-wk-old ICR mice (male:
34.7 ~ 1.24 g in body weight) in a dose of 200 mg/kg. The mice were then divided into two groups; one group was given four subcutaneous injections (each 0.l-ml dose), at 24-hr intervals, of a solvent [1$ n-propanol and 10$ (w/v) ICR
mouse serum in physiological saline] containing human G-CSF
(50,000 units/mouse) while the other group was given only the solvent in accordance with the same schedule. Four hours after the last injection, the mice in each group were _infected with Listeria monocvtogenes 46 (by courtesy of Microbiological Laboratory, Tohoku University, School of Medicine) by intravenous injection of 1.0 x 107 CFU/mouse.
The protective effect of human G-CSF' was checked by counting the number of mice which were alive 12 days after the infection.
Preparation of cell suspension Listeria monocytoaenes 46 was cultured overnight with shaking at 37°C in a Brain-Heart Infusion liquid medium (trade name of Difco). The culture was suspended in physi-ological saline.
Results 1 ~ 41 3gg i) Tests 1, 2 and 3 were conducted with the E, coli G-CSF (+VSE) polypeptide obtained in Example 16. The results are shown in Tables 12, 13 and 14.
Table 12 Effect against Pseudomonas aeruainosa Group CSF concentration Live mice/
(units/mouse/day) mice tested Solvent 0 0/10 CSF-containing solvent 25,000 6/10 CSF-containing solvent I 50,000 8/10 Table 13 Effect against Candida albicans Group CSF concentration Live mice/

(units/mouse/day) mice tested Solvent 0 0/10 CSF-containing solvent 50,000 10/10 Table 14 Effect against Liste~ia monoc~toctenes Group CSF concentration Give mice/

(units/mouse/day) mice tested Solvent 0 0/10 CSF-containing solvent 50,000 10/10 ii) Test 1 was conducted with the E~ coli G-CSF (-VSE) polypeptide obtained in Example 20. The results are shown in Table 15.

1 34~ 389 Table 15 Effect against Pseudomonas aeruginosa Group CSF concentration Live mice/

(units/mouse/day) mice tested Solvent 0 0/10 CSF-containing solvent 25,000 6/10 CSF-containing solvent ~ 50,000 8/10 iii) Test 1 was conducted with a CHO cell derived, puri-fied human G-CSF sample (+VSE) that was the same as what was used in the analysis of amino acid composition in Example 27. The results are shown~in Table 16.
Table 16 Effect against Pseudomonas aeru~inosa Group CSF concentration Live mice/

(units/mouse/day) mice tested Solvent 0 0/10 CSF-containing solvent '. 25,000 9/10 CSF-containing solvent 50,000 10/10 Substantially the same results were attained when Test 1 was conducted with a 0127 cell derived, purified human G-CSF sample which was the same as what was used in the analysis of amino acid composition in Example 27.
iv) Test 1 was conducted with a CHO cell derived, purified human G-CSF sampl a (-VSE) which was the same as what was used in the analysis of amino acid composition in Example 33. The results are shown in Table 17.

1 34~~ 389 _94_ Tab Le 17 Effect against Ps u~"~i mog~s ae,~uainosa Group CSF concentration Live mice/

(units/mouse/day) mice tested solvent 0 0/10 CSF-containing solvent 25,000 9/10 CSF-containing solvent ~ 50,000 10/10 Substantially the same results were attained when Test 1 was conducted with a 0127 cell derived, purified human G-CSF sample which was the same as what was used in the analysis of amino acid composition in Example 33.

Claims (12)

1. A DNA comprising a nucleotide sequence coding for the following amino acid sequence:
(Met)n Thr Pro Leu Gly Pro Ala Ser Ser Leu Pro Gln Ser Phe Leu Leu Lys Cys Leu Glu Gln Val Arg Lys Ile Gln Gly Asp Gly Ala Ala Leu Gln Glu Lys Leu Val Ser Glu Cys Ala Thr Tyr Lys Leu Cys His Pro Glu Glu Leu Val Leu Leu Gly His Ser Leu Gly Tle Pro Trp Ala Pro Leu Ser Ser Cys Pro Ser Gln Ala Leu Gln Leu Ala Gly Cys Leu Ser Gln Leu His Ser Gly Leu Phe Leu Tyr Gln Gly Leu Leu Gln Ala Leu Glu Gly Tle Ser Pro Glu Leu Gly Pro Thr Leu Asp Thr Leu Gln Leu Asp Val Ala Asp Phe Ala Thr Thr Ile Trp Gln Gln Met Glu Glu Leu Gly Met Ala Pro Ala Leu Gln Pro Thr Gln Gly Ala Met Pro Ala Phe Ala Ser Ala Phe Gln Arg Arg Ala Gly Gly Val Leu Val Ala Ser His Leu Gln Ser Phe Leu Glu Va1 Ser Tyr Arg Val Leu Arg His Leu Ala Gln Pro wherein n is 0 or 1.

2. A DNA according to claim 1, having the following nucleotide sequence:
(ATG)n ACC CCC CTG GGC CCT GCC AGC TCC CTG CCC
CAG AGC TTC CTG CTC AAG TGC TTA GAG CAA GTG
AGG AAG ATC CAG GGC GAT GGC GCA GCG CTC CAG
GAG AAG CTG GTG AGT GAG TGT GCC ACC TAC AAG
CTG TGC CAC CCC GAG GAG CTG GTG CTG CTC GGA
CAC TCT CTG GGC ATC CCC TGG GCT CCC CTG AGC
AGC TGC CCC AGC CAG GCC CTG CAG CTG GCA GGC
TGC TTG AGC CAA CTC CAT AGC GGC CTT TTC CTC
TAC CAG GGG CTC CTG CAG GCC CTG GAA GGG ATC
TCC CCC GAG TTG GGT CCC ACC TTG GAC ACA CTG
CAG CTG GAC GTG GCC GAC TTT GCC ACC ACC ATC
TGG CAG CAG ATG GAA GAA CTG GGA ATG GCC CCT
GCC CTG CAG CCC ACC CAG GGT GCC ATG CCG GCC
TTC GCC TCT GCT TTC CAG CGC CGG GCA GGA GGG
GTC CTG GTT GCC TCC CAT CTG CAG AGC TTC CTG
GAG GTG TCG TAC CGC GTT CTA CGC CAC CTT GCC
CAG CCC

wherein n is 0 or 1.

3. A DNA comprising a nucleotide sequence coding for the following amino acid sequence:
Met Ala Gly Pro Ala Thr Gln Ser Pro Met Lys Leu Met Ala Leu Gln Leu Leu Leu Trp His Ser Ala Leu Trp Thr Val Gln Glu Ala Thr Pro Leu Gly Pro Ala Ser Ser Leu Pro Gln Ser Phe Leu Leu Lys Cys Leu Glu Gln Val Arg Lys Ile Gln Gly Asp Gly Ala Ala Leu Gln Glu Lys Leu Val Ser Glu Cys Ala Thr Tyr Lys Leu Cys His Pra Glu Glu Leu Val Leu Leu Gly His Ser Leu Gly Ile Pro Trp Ala Pro Leu Ser Ser Cys Pra Ser Gln Ala Leu Gln Leu Ala Gly Cys Leu Ser Gln Leu His Ser Gly Leu Phe Leu Tyr Gln Gly Leu Leu Gln Ala Leu Glu Gly Ile Ser Pro Glu Leu Gly Pro Thr Leu Asp Thr Leu Gln Leu Asp Val Ala Asp Phe Ala Thr Thr Ile Trp Gln Gln Met Glu Glu Leu Gly Met Ala Pro Ala Leu Gln Pro Thr Gln Gly Ala Met Pro Ala Phe Ala Ser Ala Phe Gln Arg Arg Ala Gly Gly Val Leu Val Ala Ser His Leu Gln Ser Phe Leu Glu Val Ser Tyr Arg Val Leu Arg His Leu Ala Gln Pro wherein n is 0 or 1.

4. A DNA according to claim 3, which is a human complementary DNA.

5. A DNA according to claim 3, which is a human chromosomal DNA

6. A DNA according to claim 3, having the following nucleotide sequence:

ATG GCT GGA CCT GCC ACC CAG AGC CCC ATG AAG
CTG ATG GCC CTG CAG CTG CTG CTG TGG CAC AGT
GCA CTC TGG ACA GTG CAG GAA GCC ACC CCC CTG
GGC CCT GCC AGC TCC CTG CCC CAG AGC TTC CTG
CTC AAG TGC TTA GAG CAA GTG AGG AAG ATC CAG
GGC GAT GGC GCA GCG CTC CAG GAG AAG CTG
GTG AGT GAG TGT GCC ACC TAC AAG CTG TGC
CAG CCC GAG GAG CTG GTG CTG CTC GGA CAC TCT
CTG GGC ATC CCC TGG GCT CCC CTG AGC AGC TGC
CCC AGC CAG GCC CTG CAG CTG GCA GGC TGC TTG
AGC CAA CTC CAT AGC GGC CTT TTC CTC TAC CAG
GGG CTC CTG CAG GCC CTG GAA GGG ATC TCC CCC
GAG TTG GGT CCC ACC TTG GAC ACA CTG CAG CTG
GAC GTC GCC GAC TTT GCC ACC ACC ATC TGG CAG
CAG ATG GAA GAA CTG GGA ATG GCC CCT GCC CTG
CAG CCC ACC CAG GGT GCC ATG CCG GCC TTC GCC
TCT GCT TTC CAG CGC CGG GCA GGA GGG GTC CTG
GTT GCC TCC CAT CTG CAG AGC TTC CTG GAG GTG
TCG TAC CGC GTT CTA CGC CAC CTT GCC CAG CCC
wherein n is 0 or 1.

7. A recombinant vector containing a DNA according to claim 1 or 2.

8. A recombinant vector containing a DNA according to claim 3, 4, 5 or 6.

9. A transformant having a recombinant vector containing a DNA according to claim 1 or 2.

10. A transformant having a recombinant vector containing a DNA according to claim 3, 4, 5 or 6.

11. A process for producing a human granulocyte colony stimulating factor, which comprises culturing cells transformed with a recombinant vector containing a DNA according to claim 1 or 2, and recovering the produced human granulocyte colony stimulating factor.

12. A process for producing a human granulocyte colony stimulating factor, which comprises culturing cells transformed with a recombinant vector containing a DNA according to claim 3, 4, 5 or 6, and recovering the human granulocyte colony stimulating factor.